Thanks for listening!

Jeff

I just can’t imaging using vi(m) for development… why did they “invent” IDEs: to make the life of developers easy by automating annoying tasks. Like your import-statement-issue: why bother opening another file and copying the imports when the IDE can automatically do it for me (it’s definitely less work to add the jars to the project config than opening another file with similar imports each time i need those imports

So for me, just for quickly editing files (without any need for anything), ViM/vi or any simple text editor wins, otherwise all those IDEs win (Eclipse or Netbeans for Java, Visual Studio for C# and all the .NET stuff). Maybe there’s an exception to C++ development… it’s just not that high level as other programming languages ;P

Interesting stuff about the paperless thing, didn’t know all the OCR stuff was working so well in the of the shelf apps.

A thing that bugs me is the security of the “digital signature”. I wonder when there will be a big scandal in the news and only real digital signatures are accepted. I fear you didn’t mention anything about that. Any problems with adoption?

As you point, wicket keeps some rules to persist sessions but don’t do it always. This is the way to persist it

Quick unimportant input: You guys should re-record your outro with the microphone not directly in the line of fire of your breath. That “Twitter” hits hard.

It’s not in the downloads section.

Really? You should probably avoid the internet all together.

This whole Growl thing has left a bad taste in my mother fucking mouth.

http://code.google.com/p/growl/

To say that Google is worse about privacy than Facebook is, simply put, wrong.

Amazon AWS returns an Access Denied error message:

https://s3.amazonaws.com/bcoders/astycrapper_jeff_another_telemarketer.mp3

AccessDenied
Access Denied
E001CC2DE124DEE2FTpJHbS+8egHvHVhjpjRg+qb37IaGHLhzfRUkb4O3xAGJr77qMhNYAR7aPOOOmnD

I laughed out loud when he suggested that people could build entire mobile products using JavaFX as the UI framework. The suspension of awareness of current litigation is just unbelievable on this point. Why would anyone want to build new mobile products with Oracle’s software now? Especially Java-based software?

I was grateful that Apache’s departure from the JCP was at least given lip-service, although I understand how impossible it would have been to spend more time discussing it.

Politics aside, this was a great interview given the unfortunate limits of the day. It is terrific to hear that the Java platform is moving forward with exciting features. I hope downstream consumers of Java use these new features in a timely manner as they become available.

It might sound better to you. The only problem with Levelator is any background noise also gets brought up to the forground.

Hope that helps.

I like your podcasts, and I’m sure this one would have been interesting again, however the difference in volume between the different speakers made it impossible to listen to for me. Especially Henrik Stahl’s volume was way too low, so it was a choice of hearing Henrik by turning my volume up and getting my eardrums ripped by you guys every time you started laughing, or not hearing Henrik at all.

Keep up the good work though! Just fiddle your mixer settings a bit more next time!

Marc.

Out of the JVM based-frameworks I’m still putting my hat in with Wicket, however now I’d take a serious look at Grails.

After watching this episode and knowing you’re a Wicket fan, where and when you would use Wicket versus Angular?

I’m curious because I’m doing the painful work on choosing a JVM based web framework and would like to hear your take on using Angular. If one went with Angular, it sounds like it completely separates the web server logic from the presentation, so in the web server, all you’re doing is writing a REST or service API for the application and you could have HTML and JavaScript only developers writing the presentation layer.

Blair

Hmmm, maybe your player doesn’t support 302 redirects (or I implemented the redirect improperly). I’ve switched to 303s instead, let me know if your player works better with that.

@craig

Do you know if it’s getting this error while downloading the cast or while it’s trying to download the feed file itself? Our media is hosted on S3 now, so there *shouldn’t* be any type of bandwidth problem.

It is already some podcasts (last 4 or 5) that I cannot download from my android app ACast (all other podcasts I am subscribed to continue to work fine).

I always get a ‘Connection reset by peer’ is there something you know I could do?

The discussion about backups got my attention, because my backup system is very similar to Alex’s. I use rsync to back up onto external hard drives, and then swap the newly backed up hard drive with another at a different location. The concept of passive read errors kind of freaked me out because I’m now wondering how many of my backup files have become useless over time.

Alex, I was wondering if you would be willing to share some thoughts on detecting and addressing these passive read errors. Do you have any suggestions on the types of checks I can perform, and can you recommend any tools for doing these checks?

I think that rsync can be made to perform a checksum-based file comparison that goes beyond using timestamps and filesizes, so maybe just running rsync (much more slowly) with checksum comparisons every couple weeks would work fine, though I’m guessing you have some better suggestions.

Yes, I know I should be using Crash Plan.

Thanks and keep up the great casts!

-Nick

Bill

Putting aside the blessing/curse of one dominant vendor (of libraries/practices too!) in the .NET space, I wonder whether or not this could also be related to the fact that in many enterprises .NET is used for UI only, while services are written in Java? Unit/automated testing UI is *hard*, and is rarely ever done AFAIK – especially by developers.

I’m not looking for excuses, just wanted to add a touch of gray

Danilo Mendoza :
I consider offensive in a show like this to use the term of “retarded” to qualify an idea.
Other than that, keep on the good work.

Hodor!

I appreciate that you are trying to resolve this issue, but I hate to say it but the problem still exists
Its not in the section that you removed, those characters were properly escaped so they can be put back
In order to help you find the place where the problem is, I’ve put together a package for you with detailed instructions on where the problem is which I really can’t outline here in the comments section.
The file can be found here:
http://dl.dropbox.com/u/6042628/bcode.zip
I’ve also done a quick check on itunes and I can only see up to episode 35!
I hope this helps!

If I do a curl on our feed:

$ curl -s http://basementcoders.com/?feed=podcast | grep aitch
Craig hates the title “Software Engineer”, well apparently Chris Aitchison does too, but for different reasons. Are we Engineers or Gardeners?

You’ll see those “squares” are actually escaped properly so there shouldn’t be any problem with XML validation. Perhaps it’s just your reader that can’t deal with it. That being said, I’ll try to fix those “stylized” quotes. WordPress seems to like introducing them.

> Oracle donates Open Office to the Eclipse Foundation

s/Eclipse/Apache Software/

I enjoyed the .NET discussion because a lot of it rings true for me, though comment #2 above makes some great points. Many of the same criticisms, at least as far as developer competence goes, could be leveled at the Java community – there’s a huge base of apathetic developers using Java in a corporate setting who aren’t aware of much outside of what the boss says to use. Still, the image problem for .NET is pervasive. The .NET podcasts I listen to (good to have a rounded perspective!) perpetuate my stereotype that if even a prominent .NET developer can’t find it though File->New Project in Visual Studio, it’s not legitimate or doesn’t exist.

In short, Microsoft is doing a way better job than ever before at supporting serious developers by providing a strong development ecosystem.

IMHO, the real image problem with .net is because it is the choice of the majority of corporate business application coders who tend to be the weakest of developers. When I look at what Microsoft has been doing since releasing .Net 10 years ago and considering the user community they have to support, they have been making huge improvements.

Cheers, and keep up the podcasts!

Tommy
(occasional .Net developer)

Love your podcast but I’m having a little problem with your podcast rss feed in receiving this epp.

I’m usings a program called “Podcaster” on the Iphone (I don’t like itunes) and it comes up with the following error.
“NSXMLParseErrorDomain error 9″

I’ve run the feed through “http://validator.w3.org/feed/” and it has a few errors.

Checking the XML, There is an invalid character in the url
“http://chrisaitchison.com/2011/05/03/you-are-not-a-software-engineer”
Hex value 1c before the http://

Hope this helps.

Holger

BTW the CoffeScript thing I’ve been tinkering with is to extend CoffeeScript to add optional type annotations…

https://github.com/jstrachan/coffee-script/blob/master/TypeAnnotations.md

which lets you do optional static type checking (by passing the generated JavaScript through the Google Closure compiler in a separate build step)

However, the bit could co-variant and contra-variant is a little bit off. Using +/- doesn’t restrict the type of objects that you can put into Stack, you can still only put T’s or subclasses of T’s, the +/- tells Scala how Stack objects relate to each other. Invariant means Foo[Object] and Foo[String] have no sub-typing relationship. +T means that Stack[String] is subtype of Stack[AnyRef] and -T means Stack[AnyRef] is a subtype of Stack[String].

scala> class S1[T]
defined class S1

scala> new S1[AnyRef]: S1[String]
:7: error: type mismatch;
found : S1[AnyRef]
required: S1[String]
new S1[AnyRef]: S1[String]
^

scala> new S1[String]: S1[AnyRef]
:7: error: type mismatch;
found : S1[String]
required: S1[AnyRef]
new S1[String]: S1[AnyRef]
^

scala> class S2[-T]
defined class S2

scala> new S2[AnyRef]: S2[String]
res3: S2[String] = S2@391a74ef

scala> class S3[+T]
defined class S3

scala> new S3[String]: S3[AnyRef]
res5: S3[AnyRef] = S3@16721ee7

scala> new S3[AnyRef]: S3[String]
:7: error: type mismatch;
found : S3[AnyRef]
required: S3[String]
new S3[AnyRef]: S3[String]
^

The way I see it, assembly is the ultimate untyped language. There is nothing in the assembler to determine whether you are putting 4 characters, two 16-bit signed integers or one 32 bit unsigned integer into a register. (On a machine with 32 bit registers.) You can copy a “pointer” (or address) into the middle of a string. You can treat part of a 60-bit floating point value as a pointer or even load it and jump to it as it were code.

Most of these things I mentioned above are bad(tm) things to do though I can think of a few funky things where you treat the bits as one type to speed up working with them.

Or look at it from the point of “run time” vs “compile time”. In assembly, nothing much is checked at compile/assemble time but the failures happen at runtime whether they generate a true error or just go on running and doing unintersting things.

http://img248.imageshack.us/i/scalacheckout.png/

Let us know if you have any trouble.

Michael Galloy has a summary of his points.
http://michaelgalloy.com/2009/05/14/randal-schwartz-talk-about-dynamic-languages.html

Regarding the contracting stuff. I live in Poland and I Am 25 year old. I got my first dream job as the programmer in mid 2005 when I was 19. I worked as PHP programmer for next 2 years until I switch to Java.

From mid 2005 to end of 2008 I had 7 different positions. The longest time I spent in one company is 8 months. The shorten was position in Accenture Technology Solutions – 1 month. That was pretty funny when I was on interviews and people was asking me – wow, how old are you? You have so many positions like man with 15 year experience. Just imagine guy who 22 year old and had 7 jobs. It was me.

I was singel and I really don’t need stability. Looking for a job was something natural for me. I boring myself really fast. When you aware of code quality you and you wish to write really good code this is even simpler. When you work whole year or two on supporting some old project you dream about new project with new version of Hibernate, Spring, EJB or whatever you would like to try.

I solved this problem in December 2008 when I started my own business. Well, it wasn’t real business at all (and still isn’t). I was employed as company and paid for hours I work on project. There was few motivations to do this.

First – I didn’t like that my resume may became spaghetti with 20 positions in 2015.

Second – it is much simpler to switch project. When someone gets contractor he have some budget and time to bring working piece of software. When somebody have normal employee they don’t have to bring or prove anything. If project will fail – they are moved to another.
Simply – contractor is paid different than normal employes. Both have different goals. I would suggest that fixed budget, contractors and scrum is the best way of doing software.

Third – taxes. When you work as employee you pay 18, or 40% income tax. When you are self employed you pay 19% and have notion called “cost of income”. For example you may minus car price from your overall income. Don’t you need your car to travell to client? Sure you need even if you have permanent work. You drive car to and from work. You buy books to learn yourself or better computer with 8 GB ram to avoid Eclipse hung every 5 minutes. As normal employe you don’t have “cost of income”. You pay income tax from your overall salary. If you bought very expensive training which consumed your whole salary – your income won’t shrink.. but it should, isn’t?

Four – you won’t get 200k salary as normal employee. In Poland employment cost is 50% or your salary. Employer need to pay your insurance, contribution for retirement and so on. In fact if you earn 100k gros you employee another 50k. In fact you earn 80k net. Where is rest of the cash, where is 70k from overall cost? What is also important – I don’t belive that our goverment will guarantee us retirement. If you earn 200k per year you wish to get similar cash when you retire, isn’t? In Poland you’ll get less than a half. When you’ll die before retire children will get nothing and system will consume your money. I am not the punk but I simply don’t like system because it’s not fair.

Some people are ready to take care about own future, cash and retirement. Some one are not.

Jeff’s a born hustler; he’s good at it, too. He doesn’t have to worry about the stability; from what I can tell, he doesn’t even want it. (Good for him, incidentally. This isn’t a criticism.)

Other people are not in that position. Maybe they don’t have a problem with “always being on the hook,” so to speak (although I find being a contractor is being more responsible to demands than being an employee is.)

The key here, for me, is that it all comes down to a specific set of circumstances. I’d be afraid to say “you are a contractor at heart” or “you should stick to employment, Mr. Dudley Dull that you are.”

Interesting talk that, sadly, came across really poorly to me.

As for cost/performance, I think it would be impossible to resolve that debate. What is the value of fault tolerance including soft logic faults, being able to run full db2, redundant and hotswappable processors, database transactions per second as fast as you want, centralizing all IT in an organization, full backwards compatibility to 1960, reliable parts and service, and on and on? vs. the probability that your organizational needs will change and you’ll be somehow locked in and unable to upgrade to more appropriate tech?

Even though I don’t really care about any of those things whilst doing my personal computing, a used G5 s/390 runs about the same price as a high-end laptop these days and the inner geek in me finds that hard to resist, even with all the electricity it would consume. (I am trying to convince the wife that it could heat the house in the winter.)

To me at least, the mainframe seems to fit the centralized (also monopolist?) business model of certain other large, centralized businesses well. The biggest harm that comes from an IBM monopoly, at least from my point of view, is loss of innovation amongst mainframe makers and that other businesses are forced to move their servers towards being more mainframey. The distinctions are starting to blur, and mainframes are probably losing ground. In a more competitive environment, I think we wouldn’t even think of asking: are mainframes really better? It would be obvious they were.

I started out programming computers when I was 9 with z80 assembler on CP/M. I spent a lot of time learning whatever interested me, but never was that interested in the theory of why everything worked, and hence I didn’t really have the discipline to teach myself. I was more interested in “getting things done” if I thought it seemed cool.

When I got to Uni I studied maths and computer science. I think computer science is really a branch of applied maths. I learned the usual stuff about why some algorithms work better than others, or scale better and how to analyze them; different ways of thinking about programming; about structuring representing data and problems; specifying and verifying my designs and programs; how processors and operating systems in general work etc. etc.

I had a lot of “ah ha” moments when things I sort of knew intuitively just fell into place in my head. In addition, I developed a certain confidence that I could analyze a problem and not have the sinking suspicion that if I just knew more, there would be a much easier way to solve it. Or I could see a problem and see that it wasn’t solvable, but that there might be a good approximation to the solution. And I also got a sense of my own ignorance of the subject, and an ability to read even the most techincal computer science articles to help repair some of that ignorance. Recognizing my own incompetence was very important for me because before I had gotten a little cocky about my own abilities. I also met a lot of really cool people who shared my passion, and learned some really interesting things from them.

The fact that I had to complete assignments which wouldn’t have interested me before for whatever reason was also important, in that I got to see what other people felt was important to learn, and mostly there was a good reason for them wanting me to learn it. Left to my own devices, I wouldn’t have studied everything I did.

So all in all, even if I could have embarked on a structured series of lectures/assignments/books on my own if I had the discipline to carry it through, I still thought the experience was worthwhile, and that knowing computer science is worthwhile even if you are working in the trenches. Since then, I’ve met some amazing self-taught programmers who haven’t been to school, and know quite a bit of what I learned in computer science and often more. For them, they did just fine or better than me without a degree.

since you were kind of wondering what the SCJP says about the parameter handling in java. I had a look in the SCJP for Java 6 Study Guide and it says the following:

“When you pass an object variable into a method, you must keep in mind that you’re passing the object refrence, and not the actual object itself. [...] More importantly, you must remember that you aren’t even passing the actual reference variable, but rather a copy of the reference variable.”

This kind of explains it quite good I think and is the same what you were saying in the podcast.

Keep up the nice podcast and send my best regards to the newest member Mr. Asty Crapper

Ben.

I don’t think IBM would be eating up and shelfing companies like PSI [1] if it weren’t shit scared of the possibilities; not to mentioning threatening legal action against OSS Mainframe emulators [2]

[1] – http://www.theregister.co.uk/2008/07/02/ibm_buys_psi/
[2] – http://arstechnica.com/open-source/news/2010/04/ibm-breaks-oss-patent-promise-targets-mainframe-emulator.ars

Great podcast! I listened with interest to your discussion that “someone should write the Pet Store in each UI framework and then compare them”.

As the author of an Open Source cross-UI-framework widget, I needed to demonstrate my widget running on lots of different frameworks. And because I wanted to demonstrate my widget, not the surrounding application, I decided to use the same simple (but not trivial) example application for each one.

So I have written an almost identical app (not Pet Store, but not far off) in Android, GWT, JSF, JSP, Spring MVC, Struts, Swing and SWT. The code is readily available, well documented, and written to be easily comparable. Somebody just needs to do the analysis (of strengths and weaknesses, or lines of code, or whatever).

You can download all the examples along with my project, Metawidget, from http://metawidget.org

If you get chance to look at it, I’d be most grateful for your thoughts.

Regards,

Richard.

You are right, I’ll add that as note to correct in future revisions. Thanks for the feedback!

If you can find someone that’s done a better web framework comparison matrix and released it to the public for free, then let’s see it. I just felt like it was equivalent to someone releasing open-source software and then instead of submitting patches and improving on it, you complain about how it doesn’t work for you. You did have some constructive discussion later on which was interesting.

We’re doing a platform switch at work and Matt’s spreadsheet was really helpful for us to talk through the various frameworks and issues. The point, at least to me, is to take it, strip out the frameworks you’re not interested in, change the ratings you disagree with, and see where YOU come out. It’s not to view Matt’s word as gospel.

FWIW I really enjoy the podcast and you guys do great work with it by taking the time and making it freely available so keep up the good work.

http://raibledesigns.com/rd/entry/how_i_calculated_ratings_for

When you where explaining “object” I missed again the term: singleton, because that’s what it is. If you think about it in a OOP Language terms like class and object make need to be distinguished and in my opinion scala does that very well.

Summing up I really like your chosen scala topics and how you explained them (that and the presentation style + teen wolf reference got me ).

regards andreas

Would like to hear your thoughts though on why you might feel this spreadsheet should be considered useful. Perhaps take Matt out of the equation, and let’s pretend *I* produced it instead. Would I not be held to account for how the metrics were gathered? Or should everyone just take my word for it?

Maybe I reacted too closely to this: http://xkcd.com/386/

But the article (and comments) over at Peter’s blog pretty much sums it up for me (and by “me” I means Craig, as I don’t speak for all the Basement Coders which is evident in listening to the podcast) : http://ptrthomas.wordpress.com/2010/12/04/comparing-jvm-web-frameworks-a-response-to-matt-raible/

I generally enjoy the podcast but in this episode I was saddened. You could have had an interesting discussion about the merits of various web framework but instead you completely miss the point of the matrix and wasted our time by cutting down a fellow developer and discouraging others from sharing their opinions with the community at large.

In summary: where’s my flying car?

I sometimes have problems with the Vimeo player (on Mac). I find if I click on the “Vimeo” text in the lower right portion of the player which takes you directly to the video on their site it usually works there. If not, restarting your browser might help

Included “free license for JRebel ($60 value!) to the first 20 people who pledge $20 $10 or more” in the message.

Good luck and Happy Movember!

Thanks,

Gary Blomquist

I just donated but forgot to put JRebel in the message! Can you please include me.

Many thanks and all the best with your fund raising.

Jon Hill

Regards,
Benjamin Peter

Looking forward to hear more of your episodes.

iBatis, actually they are going through a bit of transition where they are forking the project outside of Apache and continuing work, this link you might find interesting:

http://basementcoders.com/2010/06/a-case-for-ibatis/

thanks for the episode. I think the chatting at the beginning is okay, it makes the cast kind of special and, as you said, like some buddys talking in a nice round.

But I have a question, what was the other DB access library you mentioned, at the beginning of the podcast, as a complement to hibernate. “ibdus” or something like that?

Thanks for your answer.

This is much much more efficient:

find . -name \*.scala | xargs grep ‘def from’

and if you just want a list of files that contain ‘def from’ use:

grep -r -l ‘def from’ *

e.g.
Invoice #0001
Customer
Date
Line Item 1 – Parking Ticket $250
* Parking in a Handicap Zone $200
* Expired Meter – $50
Line Item 2 – Sewer Line Maintenance $1500
* Excavation – $500
* Sewer Pipe Repare – $1000
Payment Details
etc…

After that you’d need to design an interface for a Cash Register

eg are you using git pull etc to keep the remote git stuff in sync etc.

Thanks.

I actually listened to the ads just to see how long they went on for. 3 minutes and 13 secs. I did not listen to the interview.

In case it helps, I listen to podcasts while I’m doing something else, usually driving to work, mowing the lawn, fixing something around the house. If 10% of a 30 minute activity is spent listening to something I do *not* care about, I’d rather just listen to my own thoughts.

I understand the need for sponsorship, but this is not 1999. Look at other podcasts publishers for ideas on building loyalty and tying it to revenue. I’m sorry, but I won’t be back.

Best of luck.

…aaand this is how I know America is finished.

Actually it wasn’t a hand held device at play, it was two condenser mics, two MBPs and Garage Band in a *very* noisy venue with a soft spoken guest Sorry we couldn’t clean it up anymore. Near the end of the cast you can actually detect the drift (echo) between the two MBPs recordings… oops

Can I just say one thing, and it might have just been my computer because Guillermo’s was fine, but Garage Band cannot be trusted on its own for live recordings. You should always record in tandem with someone else. It crapped out on my computer twice. Once was at the end of this interview, luckily though the underlying recording was recoverable from the .band package (although the track wasn’t showing up in GB, can you say… HEART ATTACK?)

If anyone has any tricks to get rid of the background noise, please let us know over at The Basement Coders!
<snip>

Gordie: “I’m from Canada, they think I’m slow eh?”

The great thing about digital media is it’s RAM (unless you stream it), So skip past the commercials if they bug you. The companies we have sponsoring us either help the podcast out in some way (i.e. The Server Side hosting our podcast file so we aren’t pwnd) or help us get the word out to their own user base so we can grow our audience.

Believe me when I say this, we aren’t making a cent, we do this for fun. That being said, we’d love to be able to do this full-time, but I don’t think you are allowed to have this much fun at work

Re-run from What’s Happening just called, he needs his tape recorder back for the Doobie Brothers concert tonight.

Amazon C3 is cool -> Amazon EC3 is cool

The end continues! (67:50)
Moderator/Random guy:
About cheapest you’re going to get is $70 a month and for a lot of people that’s not trivial
(inaudible stuff) Moderator names people?
JG: Geeky room my friends. Thanks, appreciate it.

“But we ended up concentrating on Java code” -> But we oddly ended up contentrating on Java code

“cosign” -> cosine typo, several times. Also “Sine” should always be lowercase.
gimp, blender, openSSO -> initial cap for each of these.
“people doing desktop software and the original versions of the Java store” -> (needs an em-dash, he changes the subject) people doing desktop software — and actually the original versions of the Java store
“get the Java store down at Sun” -> get the Java store done at Sun
“would support not just support Java but all sorts of code” -> would not support not just Java but any kind of code (inaudible)
“gird” -> grid typo
“talking to SSN folks” -> talking to the ForthRock folks (based on Wikipedia’s OpenSSO article) (63:50)
LinkiedIN, Framville, Zinga -> LinkedIn, Farmville, Zynga typos.

Minor:

“Cagary” -> Calgary
“Master’s as well” -> Master’s as well from CMU
“and you know, did a wide variety of consulting jobs” -> he doesn’t say “you know” here.
“stuipidist” -> typo
“working on this project that had people in California, Texas and Connecticut” ->
people in San Jose, California; in Austin Texas; in [garbled], Connecticut.
“lasted a year” -> lasted a year, year and a half. (maybe unnecessary correction?)
“and so he was actually trying to join Sun at the same time” -> and so he was trying to get me to join Sun

@Cameron, don’t be hard on yourself, I could barely put together sentences after JavaOne.

“gustoes” -> “good stewards”
“Sandra and Baltimore” -> “Cassandra and Voldemort”
“theorem approval” -> “theorem proving”

“Sun’s board was controlled by an extremely small number of institutional investors, and so it was really those institutional investors who were driving everything. The sale had nothing to do with the business or what would be good for the employees or any of that. It was totally a bunch of investment banks needing liquidity now. And they were looking for the best terms with the most certainty.”

Oracle therefore thinks that SUN must be doing it all wrong, so Oracle stepped up and made decisions for SUN.

Oracle has every rights to make decisions for SUN. After all, it owns SUN.

What Oracle had done to SUN’s senior staff might be immoral, but perfectly legal.

Oracle was simply doing what it does best…make money “at any cost”.

Does Oracle owe Gosling anything ?

Does Gosling think the world owes him anything ?

Sorry. Just another quote I had to share, just in case someone didn’t get to the 23rd minute.

That’s a great quote!

Nice choice with the Droid X. Once you root it and play with the video on it. It is really nice. I also got the dock with HDMI, which works very well.

Welcome to the club.

SC

The other type that really bugs me is the guy/gal who thinks their needs are more important than whatever it is you might be doing at the moment. I had a guy the other day who needed help setting up a project. I was in the middle of bug fixes and he didn’t care, he insisted his environment be setup ASAP. I think that bugs me because I am always respectful of other people’s time, especially when they are helping me out…

I had to deal with few of them on opensource projects.

At a point I was stuck and needed a working case to test the modifications that I did. I asked few times and never been able to get a example from the guy.

Look like he wanted me to complete the feature all alone without his help. I hate that kind of guys. I took my free time to help him, but he can’t take few minutes for me.

Thanks for this great podcast. I have added a link to it from the Camel articles site (our link collection)
https://cwiki.apache.org/confluence/display/CAMEL/Articles

Will take some hours to sync to the static html pages
http://camel.apache.org/articles.html

-
Claus Ibsen
Apache Camel committer

Also, someone get Craig a whiteboard so he can keep track of the episode numbers He’s not even into double digits yet and he’s lost track.

Regards
Noel Alex,
Java Developer

IS the Mainframe a replacement for “Cloud”? No, but a mainframe can/should be part of any Cloud initiative, simply because it’s already there. Parts of an internal IaaS (Infrastructure as a Service) initiative could very easily be implemented on Linux Servers running on a Mainframe. And since the Mainframe still holds anywhere between 50-70 of the Corporate data at many companies, the Mainframe MUST be part of any Cloud initiative that integrates with, or uses existing data.
For more info on Mainframe & Cloud: have a look here: http://bit.ly/9i4jRJ

Disadvantages of server based technology:
- The need to “system support” many boxes each of which seems a little different.
- Regaardless of system image installation it awalys seems that fighting conflicts especially because of system updates is a concern. I have seen 2 weeks to try and track down a DLL conflict.

Lessons mainframe could learn:
- Use the proper tool for the proper job.
- Use mainframe based Java and XML packages to take care of the boiler plate activities.
- Just because a sollution worked in the past, doesn’t mean there isn’t a newer technology that could be leveraged more effectively.
- Mainframe programmers need to learn not to be so protective of their boundaries. This is partially due to unfamiliarity and frustration with the seeming unstable server development software.

Trends:
- There is a beginning trend to run virtual servers on mainframe platforms. This promises to have the best of both worlds: flexability of of development plus stableness of the system environment.

For instance, don’t use ASP.NET MVC at work? Why not toy around with it at home? Or ask for permission at work to do a proof of concept? Your company really won’t care if the tech you are using is super outdated, so long as it works and can still be supported. Hence how you can get stuck in a technology rut. If you could show them that “hey using this new thing will make it way easier to maintain going forward” they might be prone to take your advice.

I tried to utilize this but I did not find a case for it. In 1.4, even with stateless pages, I could not see that getSession().bind(); was required anywhere. Could you show us a testcase that demonstrates where this code is advantageous?

It’s pretty bold of you to assume you have all the information on this incident, which you obviously do not. It’s always easier to jump to conclusions, but as it turns out, I think the community would agree with me on the fact that there was certainly a level of disrespect going on here. An act a CEO should never commit, yet Jason seems to do quite frequently.

Thanks for your input though.

- Nishant

I also don’t know what is finicky about gitosis? I’ve been using it for a long time and never had a problem.

for the moment i started moving some of my stuff over to github ( timvoet.github.com/DatasetCreator ) at least for the stuff that isn’t confidential.

I was thinking along the same lines for groups, it just becomes alot of setup and maintenance, but might be worth it.

i’ll let you know which route i decide to follow.

I too looked into gitosis by following ( http://scie.nti.st/2007/11/14/hosting-git-repositories-the-easy-and-secure-way ) but couldn’t quite bring myself to going through all the trouble for my limited audience of developers

how do you ( or do you have the requirement ) to manage access on a repo by repo basis. I have git repos i like to share with some ( as in collaborators or some of my freelance clients ) but don’t want those repos shared with all the people that have access to my repos.

I’ve used gitosis before, but it can get quite finicky sometimes.

any ideas?

I love your comment ” I’m not being a jerk here, I love Jeff and Joel, I just recognize their ignorance for what it is.”

My comrade recently posted an excellent analogy on why testing is important: http://app.arat.us/blog/?p=159

I think this is a classic case of forming an opinion on something based on never using it (or using it 7 years ago). Dell would not be shipping machines with Ubuntu on it if they thought their switch board would be overwhelmed with basic questions like this.

Now, I think Alexs should perhaps install Windows 7 in a VM or soemthing and kick the tires and then do an update post with his findings. Maybe his mind will change on some issues? His mind certainly will not change with Vista, to me it’s just XP with a slightly more fancy UI (every see OSX? Compiz? They kick Aeros’ ass). It pisses me off to no end that they basically used DirectX 10 to sell the bloody thing.

I don’t know much of what you are saying. But, as a long time windows user, Here are the most used and required things in windows or any other operating system for a normal user.
1. Able to navigate the file system effort lessly
2. Able to learn to create/modify/delete text or word like files
3. Able to listen songs and watch movies
4. Able to use a internet browser .

All these things can be done in a simple ,user friendly and understandable way. Learning curve for doing these things windows takes very less time than in any other unix system.

For a home user like me who knows a little bit about computers, I need
1. to be able to install and uninstall operating system
2. to be able create/delete partitions
3. to be able to install and uninstall any other windows based apps.
4. Able to connect to internet/setting up internet connection
All these four things are easy to learn and do on any windows operating system

These are things that I like in windows. Lot of unix operating systems like ubuntu ,fedora still talk about
command line options. I don’t know when these unix sycophants will try to know and understand the problems of common user. We don’t need a top class operating system. We just need a easily usable operating system. One of colleagues , who is a software engineer, unable to set up internet connection on his ubuntu. Need I say anymore? Don’t make any comments about his sofware skills. But, understand how difficult the unix systems are. If you want evidence of importance trouble free / easier installation of operating system, I will provide you the link to interview from none other than linus torvald(http://news.oreilly.com/2008/07/linux-torvalds-on-linux-distri.html).
Market share of windows speaks itself for the great usability of windows.
I still don’t know how many more decades or years , your unix guys will take to achieve the usability of common user. If you want market share, just do the above things.

What? Windows lacks, out of the box, drivers for countless network and wireless cards. Yes, I could visit the manufactures site, and do I, my point is, I shouldn’t have to and non-technical users aren’t going to have a clue (they go pay Geek Squad to do it).

I’m comparing the install processes of Linux and Windows, because Windows bloggers criticize the Linux install process. But yeah, OEM version is great. My Dell Mini 9 came with Ubuntu installed and configured. Works great. The fact that (as you’re implying) that MS cut corners because it’s just for OEMs is just silly. MS does sell retail copies of Windows, which users do use to update. Corporate users do their own installs and configuration, etc.

Yes, the updates are obtrusive. Windows Genuine Advantage? That’s obtrusive. Lets ask the clueless user if they should install software which has no end user function. Did they say no? That’s ok, we’ll ask again later.

Do you get that reboot prompt that automatically clicks “yes” after 5 minutes or something? Because that’s the one that reboots my computers “for me”, ignoring the fact I was doing some non-interactive work on the machine.

Apps install their own system libraries all the time.

Just because you have a large market doesn’t mean you should require a virus scanner. Linux has a huge slice of the server market, yet you don’t see those falling over all the time.

But my apt repositories are cryptographically signed. Training users to download random .exe files from random google searches is just waiting to be pwned. You can’t do it that easily on Linux, as to install software you either have to a) chmod a+x the download (cryptic) b) install the package, which requires a password (which sets off warning bells right there)

OEMs do NOT take care of installing basic software. They install their own flavour of spamware on the machine. This OEM practice is a point against Windows, not for it. But I don’t blame MS for what OEMs do.

RDP is not an ssh client. Explorer is lame at best for features you listed.

When you try to play a DivX file in Ubuntu, it’ll pop up some dialogs. Click OK a few times and it’ll “do what it needs to” and presto, you’re watching your movie.

Notepad could easily support UNIX line ending, MS chooses not to. Users don’t know what UNIX line endings are, they just ask me why some file they found on the network drive “looks funny” in notepad.

Users don’t need to know what EXIF is, they just want their software to automatically rotate picture as needed. The default Windows viewer ignores this information, and users everywhere have been unnecessarily rotating pictures ever since.

Windows is ass for development or as a power user’s platforms. I’m not even going to go there.

“Because OS X is Unix, Windows isn’t? It’s a stupid complain and you know it. Why doesn’t Linux have MS tools or support for them? Apple does! Besides, you want your Unix Utilities? Microsoft did that, too. See SFU and SUA.”

SFU? SUA? You talk as if those are common solutions, actually work, are maintained and don’t cost an arm and a leg. Don’t point at solutions people don’t really use and can’t, especially for SOHO/medium sized shops. Microsoft can include a decent shell and openssh tomorrow if they wanted to, but they don’t. They don’t want to play nice with other systems at all, ever. It’s not like Windows is so fundamentally different that you can’t include common tools on it, come on. If little Apple could do it, then so can MS.

Thanks for your input. I’ll reply to a few of your points.

Yes, if Linux lacks a piece of software that you require then you can’t use it. This I admitted in the article. However, for more and more people it’s fulfilling their needs. I deploy Ubuntu LTS + firefox + openoffice + auto updates for many family members, because it makes for a worry free set-up.

People who build their own systems or systematically reinstall their systems aren’t normal users, they’re power users. Normal users throw their malware infested machine out and buy another one, or get their local geek to fix it up.

My comparisons where for a retail version of XP installed on a plain computer. Customized OEM versions of XP better be easy, but that goes without saying. Heck, my Dell Mini with factory Ubuntu has a slick restore partition too.

Windows still does a poor job of driver support out of the box. This was my point. I don’t care about who’s fault it is (and to be frank, I think MS could think of some kind of solution here if they really wanted to).

The problem with many of your apologies for Windows is that you think because that’s how it’s done, that that is the only way. MS, if they wanted too, could really think outside the box on how to stream line a lot of this stuff. There are many options here, I can probably do a whole technical post on that alone, but suffice to say, MS simply doesn’t want to fix these things.

Yes, there is always 3rd party software to fix everything in Windows – that’s saving grace. The issue is that new users don’t know what stuff is good, what stuff has malware, and what stuff simply isn’t needed.

Yes, my notes on SSH and stuff are only applicable to power users/admins/programmers, but you’d be surprised how useful those tools are. I set-up my wife’s machine to auto-backup via ssh. The 2 other basement coders here both bought Macs to do development on because of its powerful UNIXy shell (and other features). Technical users often influence their peers, so wooing them is beneficial.

That’s just it. They do provide a spectacular development environment and optional developer tools, the former out of the box. .NET and MSDN make it easy to develop for Windows, before then, win32 made it “easy”. That’s how you lure third parties to develop for you platform – standardized, contral frameworks which not only result in muchly accelerated application development, but also offer free system integration, and easy deployment.

If anything, their biggest shortcoming in this regard is that the base system and associated API/ABI are TOO stable. Applications that haven’t been modified in 15 years will still run on a modern Windows system, and with the same ease of installation as before – can you say the same about pre-1.x apps on Linux 2.6 systems? I didn’t think so.

Further, there are spectacular developer tools readily available (visual studio express) for free, and pro developer tools also readily available (Visual Studio) at a cost (unless you’re a student, see DreamSpark). Of course, these are geared toward the Windows developer environment, which is understandable, I think. But if you’re hellbent on using Python, Ruby, Perl, PHP, Lisp, Java, Fortran, even COBOL or any other language under the sun, they’ve all been ported to the .NET CLR, and can be used in either managed or unmanaged mode.

I’d say, comparatively, Linux is a much more hostile environment for developers, with the complete lack of a standard base system, the LSB initiative is joke nobody takes seriously, and the base system fragmentation is patently absurd, for example, see here 18 binaries for x86_64, 31 for PPC, and 65 for x86- for a total of 114 binaries, per supported release version. So very, very friendly.

There’s intentionally unstable, ever-changing kernel-level ABI and API (see GregHK – and let it be known that since FreeBSD is able to retain ABI compatibility not only with several older versions of itself, but with Linux as well, GregHK’s claim that maintaining a stable Kernel ABI/API in Linux is impossible is a lie – it’s a failed attempt to strong-arm manufaturers into releasing the source, while lacking the leverage, or significance to pull it off). See the VMWare debacle where incremental point releases x.y.z -> x.y.z+1, break VMWare because the hooks it was plugging into were removed.

How about the mess of OTOH more than half a dozen (qt, gtk1 gtk2, motif, openmotif, tcltk, efl, wxwidgets) toolkits to choose from? Or the dozen or so competing, sometimes overlapping, and always changing sound (OTOH – ARTS, Phonon, Esound, esd, Jack, Alsa, SDL, portaudio, pulse, and of course, the Unix standard: OSS) subsystems?

Right, because having singular, centralized, STABLE framework (DirectX/.NET) to handle everything, is so much more hostile. What were they thinking, making the developer’s work easier? To use your own quip, Apple does this, too (Cocoa/Carbon/CoreFrameworks, and optionally Posix and X11, but let’s be honest, non-cocoa apps don’t count) why can’t Linux?

In fact, everyone else in the OS market, even in Unix and free software does all this to a certain degree (stable, standard base systems, at the bare minimum) why, again to use your quip, does Linux have to be so drastically different?

And your other complaints are just too ridiculous to comment.

XP has been making MS sales/money to this day. MS also updates XP continuously over the years (IE8, wifi support, SATA support, etc) via its very large service packs. It’s not like MS released XP 8 years ago and haven’t touched it since – it’s continuously updated software.

Windows 7 hasn’t been officially released yet, and thus for all practical purposes doesn’t exist.

Windows 7 is new technology and it is ready and prepared for the desktop that’s what I mean.

Good thing for them, users aren’t their target market, OEMs are.

Lacks drivers and poor auto-detection of hardware, often leaves the computer installed without network, sound or 3D video card support.

That’s what the driver cache is for. baring that, it’s what slipstreamed install media is for, and again, OEMs. But let’s be honest here, how often is “often”? Often, providing the network is up and running (it’s always been configured during the install, unless you do something ‘clever’ and go with the default (dhcp) when you know that you aren’t running it, a simple trip to automatic updates results in the appropriate drivers being installed. Unless you tuened that off, too.

Have to search the backwater download areas of various sites, running random (and dangerous) executable installers hoping that I have the right one for my particular hardware adapter.

Gods forbid you visit the device manufacturer’s site.

Lots of reboots between driver installed wasting boatloads of time

Because you can’t, oh, I duuno, drag that “windows is going to reboot now” dialog off screen and ignore it until you’re done doing what you were doing.

No live CD

Why would you need one, insofar as the installation goes? The whole LiveCD installation thing is way overdone, imo. It was neat back when Knoppix was relevant, and it has its uses as a rescue CD, but come on. Besides, you (should) know as well as anyone that the install process is designed for Windows’ target market: OEMs.

Auto upgrades seems slow. Asks a lot of questions, and is very intrusive.

Express or Custom. Uncheck what you don’t want. Press OK. Yeah, so many obtrusive questions!

No nice (central, Internet) package management system.

Updater sometimes reboots the computer automatically, abruptly terminating any running programs that the user may have delivery left running.

Windows update? I’ve always gotten a prompt (that I’ve always ignored until I felt it was time to reboot).

No central upgrading system for 3rd party software. Why doesn’t the add/remove program window in control do this?

Because it isn’t feasible, and 99% of the time, you can set an option for any given application to check for updates only when you tell it to.

First party updates tend to be handled by AU.

Upgrades can easily break stuff (no proper dependency tracking)

Dependency handling isn’t necessary because no shared libraries are being replaced, unless of course, you’re trying to install everything into the same mega-directory.

Apps install their libraries into their own folders, where’s the conflict?

as upgrades may not just include version patches, but completely new versions of software. Not user friendly at all.

That’s why you’re provided with a list of which updates are being installed, are given the option not to install them, and given enough information to look u what they do.

Frankly I’ve seen a lot more upgrades gone awry via apt-get dist-upgrade or emerge world than I have from automatic updates.

Requires virus scanner.

Yeah, sort of comes with the territory, huge chunk of the market, very tasty target.

Base operating system demands it, yet doesn’t provide one.

They’d be taken to Anti-Trust court over it and you know it. Though Windows Defender has been out for a while now (they just can’t bundle it). Again, that’s where a) OEMs come in, and b) slipstreamed install media.

Can be disabled though. Virus scanner makes computer much slower.

Right, because everything is as stupidly bloated as Norton.

especially since many required utilisation and software packages come from the Internet via the web browser as executables – not the safest delivery vector.

Just as you trust your distributions depositories, people trust the vendors who provide the product installers. Really, it doesn’t take a brain surgeon to dot download arbitrary installers from arbitrary third parties. There’s a lot of incentive for a reputable company to not fuck over their users – theit bottom line depends on it, for one.

Some kind of chmod +x system should be required – software should come in packages and/or package repositories.

Let’s not mention things like fedora repos being compromised, or a certain Debian package maintainer shipping an OpenSSH with the random number generator disabled. No, repositories would never allow their users to be left vulnerable!

In either case, there’s a certain level of trust, and using common sense involved.

There is some kind of signing system for executables, but the user can (too) easily dismiss the queries/warnings, rendering the system ineffective.

This is true of any system. You can just as easily force an install on a Unix system.

Lacks lots of basic software. PDF viewer,

OEMs take care of this.

ssh clients

See RDP and Hyperterm.

GUI ftp client

See explorer.

decent zip-client,

Support for zip is built in. See Explorer.

CD/DVD burning software,

See Explorer.

CD-to-MP3 ripping, etc.

See WMP. Well, CD to WMA, anyway.

Some stuff can be downloaded from Microsoft’s site, but this requires knowing about it, finding it and other hassles. Sometimes add-on software can wreck havoc, like after-market firewall or wifi network management software. New users would need the advice of experienced users to avoid installing malware, who have set up realistic-looking mock sites to trap and ensnare users.

Right, because nobody ever needs help or advice from experienced users when it comes to Linux. Ever. And again, a certain level of common sense is required.

Lacking lots of codecs out of the box (like DivX) with no obvious method of getting them. Media player won’t rip to MP3 by default, insisting on WMA, even though MP3 is the de-facto standard (to be fair, iTunes is guilty of this too)

To be fair, many Linux distributions don’t include mp3 out of the box, either. The obvious method of getting codecs, is to look for them online. Chances are, if you know what DivX is, you know where to get it, and there’s no real need to know until you come upon the format, usually on a webpage, and usually next to a helpful link to the codec.

The basic software that is there is not polished. The picture viewer doesn’t do EXIF rotation. Notepad doesn’t support UNIX or mac line endings. WebDAV client support is just broken.

Motepad has no need to support Unix line endings, it’s a) meant to provide basic text editing functionality, and b) if you know what Unix line endings are, you should stop pretending that you’re the “easily confused novice” and that you don’t know where and how to easily acquire an editor that meets your needs.

Exif rotating? Again, quit playing dumb, if you know what exif is, than you know where to find a viewer that does it.

WebDav? You’re obviously not talking about novices anymore.

Also, if you’re hellbent on Unix compatibility and Unix features and still want to or need to use Windows, chances are you know about SFU and SUA, regardless.

Very unfriendly environment to developers or system admins (and other such power users)

Unfriendly to such people accustomed to Unix, no doubt (Hint: They know about SFU and SUA/ Hint 2: Windows isn’t Unix, MS got out of the unix business almost 30 years ago).

Windows amins however, are spoiled with a wealth of polished, functional admin tools. Developers are spoiled rotten with a spectacular compiler, debugger, integrated, ventral framework (.Net), stable APIs and ABIs, spectacular documentation (via MSDN) rivaling of not bettering FreeBSD’s legendary handbook, and second-to-none developer tools (see Visual Studio, or even its lesser offspring, the VSExpress, also, see the DreamSpark program). Same holds true for the power users and the myriad of of powertools, toys and the customization scene.

Poor command line (shell), no ssh system,

See Powershell.

non standard build tools (like make),

You’re expecting to do Unix development on Windows? (see SFU and SUA, they take you into account, too). There’s no good reason to integrate Make into Windows. It may be standard on Unix, but once again, it isn’t Unix. Unix standards aren’t standards on Windows. You have a wealth of spectacular dev tools at your disposal, use them!

no scripting languages out of the box,

See WSH.

lacks many basic command line utilities.

Which is supplemented by the GUI utilities. They’re easy to add in, though, and if you really want Unix CLI stuff, there’s SUA/SFU.

drastically different
Yes, that would be the point of exiting the Unix market three decades ago, wouldn’t it?

in this respect. This results in Windows being a difficult platform to use when working with many open source projects or languages.

Because it’s oh so difficult to install your non-standard tools on the platform. Talk to any .NET/Windows developer and they’ll voice the same complaints about Unix.

Developers pretty much have to be using full IDEs and packages, but even in those cases, assembling libraries and build systems is still a pain with lots of trial and error.

See Visual Studio.

Cygwin is a half-solution at best.

A Microsoft sanctioned shell

See SFU/SUA and Powershell.

ssh client/server

See SFU/SUA /RDP.

and collection of utilities would be best.

See c:\Windows\System32\

Apple seemed to have done it, why can’t Microsoft?

Because OS X is Unix, Windows isn’t? It’s a stupid complain and you know it. Why doesn’t Linux have MS tools or support for them? Apple does! Besides, you want your Unix Utilities? Microsoft did that, too. See SFU and SUA.

Not very customizable or tunable by hardware makers, making it an ill fated choice for netbooks. Depending on how the netbook market goes, this may or may not resolve itself. If netbooks get more powerful for the same price, then they can more easily handle Microsoft’s operating systems. However, if they get cheaper by using even lower cost components (ARM CPUs, for example), then Microsoft’s operating system will be completely shut out.

The article’s date says 2009, are you sure it’s not 1999 in your head? See WindowsCE and NVidia abandoning Android for their ARMbook in favour of it. Also see shared source.

People don’t so much think Windows is easy to use as they’re just used to its crapulence and don’t want to bother with anything else.

It’s pretty hilarious that you come off exactly as one of those “whah whah whaaambulance Linux isn’t exactly like Windows” crybabies except just as whiney, and coming from the other side. Your whole article is a drawn out rant about how Windows lacks package management it doesn’t really need and isn’t a carbon copy of Linux, the two or three cogent points you had were lost in the deluge of whiney banter.

Suffice to say, there will be no distinct “year of the Linux desktop”. It’ll be a gradual process, as Linux slowly grows into markets

Suffice to say, that’s wishful thinking. 20 years and Linux hasn’t managed to stay above the 1% mark in terms of overall usage. Microsoft had established itself as the market leader, while coming from behind in far, far less time, twice. First with Xenix to dominate the Unix market, then with DOS and Windows. Apple managed to crack the 10% mark. Commodore managed a much more impressive figures and completely vanish is less time. I’d venture to say, even the Amiga and Atari ST managed to crack the 1% barrier.

The lack of third party software support is a huge weakness for Linux. It does not matter who’s “fault” it is. If I buy a hydrogen-fueled car and my local environment changes so that the nearest hydrogen station is 100 miles away, the car is much less useful. The how or why is irrelevant.

This is the first time I’ve ever heard a Linux user state “no one ever has to install Windows”! Usually they complain about having to re-install it frequently due to DLL-hell problems. Personally I reinstall every few months to clean out the garbage that accumulated and to load newer versions of software and drivers. I’ve also installed Linux. I don’t think anyone can claim a Linux install is as easy for a lay-person as the Windows install. Additionally, people who build their own systems manually install Windows on it, and it wouldn’t be too hard to consider that far more of those individuals install Windows as opposed to Linux, meaning there are still many more people installing Windows than Linux on an absolute level.

Windows does not have a poor installer. It is streamlined and easy to use. If you buy a ready-made system, it is designed to be customized by system builders to have software and drivers pre-installed along with the core installation. If you do an install of Windows on a pre-fab system without the restore CDs the manufacturers provide, it is up to you to gather the device drivers ahead of time. This is not the fault of Windows, but the third-parties who customize their hardware. If you build your own PC or have purchased hardware upgrades, you have drivers for your hardware ready to install. But the dramatic embellishment of how “scary” it is to download drivers on the Evil InterWebs makes for good copy I suppose. Have you seen the size of the driver lists for Windows hardware such as network cards? I think they do a pretty damn good job.

Reboots for driver installs is a genuine flaw, but considering how infrequent driver installs are, it is negligible. “Boatloads of time”? How much hardware do you have in your box? I can install all of my extra drivers in less than 20 minutes. Unfriendly to non-Windows partitions is another flaw, but Windows can occasionally be unfriendly to it’s own partitions! There is very little consumer demand for it to be friendly to non-Windows partitions.

Auto-upgrades “seem” slow is pure conjecture. Intrusive, perhaps, I don’t care, but I suppose some people do. The nag screens are there to help and advise non-experts, of which there are many. Add/Remove programs should be cleaner, but problems usually happen due to the routine being poorly written by the third party provider. If you hold to your implication that poor software availability is not Linux’s “fault”, then this is not Window’s “fault” either. Auto-reboots for updates can be shut off if you look in the options. Otherwise, it generally TELLS you that the system may automatically restart before you even begin the update, should you choose to read the display text.

A third-party software update system means Microsoft has to manage it. If you took a look at how much Windows-based software existed you would see this is not only infeasible, but logistically impossible. MS could create an auto-update interface for third parties to use, but as developers often have their own automatic updates (which likely have additional purposes other than the altruistic act of keeping the software up to date) and probably wouldn’t use it, what’s the point? The rest of your complaints about software upgrades are also the responsibility of third parties – again, not Window’s “fault” in your parlance.

Windows “requires” a virus scanner because it is used on 90% of the world’s PCs. The OS doesn’t “demand” it. But yes the security manager recommends you have one. This is the same thing as telling people to wear their seat belt when they drive. The Mac has seen a virus recently. The only reason Mac and Linux users do not worry about viruses is because there aren’t enough people who care about infecting their platforms.

There is no “nice” repository for software because there are hundreds of thousands, maybe millions, of applications available. Would you like to pay for hosting that?

Default file formats are often in competition with Microsoft file formats, so yes unfortunately Windows doesn’t support them out of the box. Some of your listed “basic” software cannot be considered basic. How many typical Windows users need SSH? C’mon now. Having “decent” default versions of much of that software is problematic for several reasons. One, MS has to pay for development of said features, and even if you don’t like them, you pay for it with the cost of the OS. Second, it opens MS up to antitrust complaints. Providing simple apps and letting users get better ones on their own based on their needs is the only reasonable solution.

The OS provides codex that the average user will use. At one point Microsoft was being litigated against due to the inclusion of the mp3 codec in Windows Media Player so they removed support for it. They have since resolved that and Vista has mp3 support once more. I installed a Linux distro a couple years ago that didn’t have an mp3 player. Once I tried to find one and figured out the poor selection and the hoops I’d have to jump through to install it, I gave up.

“Polished” apps, to you, means “doesn’t support the features I want”. That’s the beauty of Windows – find an app that does. There are dozens or hundreds to choose from.

Unfriendly to developers? Complaining there are no standard programming tools is absolutely silly. How many Windows users are developers? 1%? Why would MS spend money to develop those tools for 1% of it’s user base, 95% of which are going to want to use an alternate full-featured IDE instead.

Your article intended to prove the postulate “Windows, on the surface, seems like it would be a complete failure as a desktop operating system”. If, to you, not having SSH or developer tools contributes to this point, you clearly do not understand what a desktop operating system is for 99% of the population. Windows has many true flaws, but few of them were mentioned here.

@Isaac thanks. Can you hook me up with your awesome copy of XP, because if yours fixed all the issues in mine, I’m all for it

For a developer, unless you are doing .NET or Windows specific programming, I would say Mac OSX or Linux is the only way to go. Developing on Windows is kinda like putting lipstick on a pig. Why do I have to install MSYS or Cygwin just to get a decent shell with grep, find, xrargs…. Apparently Windows 7 shell is supposed to be “all that and a bag of chips”. We’ll see….

Poor Installer – Windows 7 Install is very very, easy don’t know what he’s going off of.

AutoUpgrader Slow – Depends on system hardware, it’s fast on my 2 GB of RAM laptop, it’s very easy to use and I can’t see it scaring anybody away.

No Third party support with Uninstalling and upgrading – is pretty simple and you can update third party software, In Vista and 7 you can choose to upgrade with third party softwares through the autoupdater, in fact Crysis Wars updates through the Windows Update.

No Virus Scanner – If they included an Anti-Virus scanner with Windows then Microsoft would get sued by symantec. I’m running the beta of Microsoft Security Essentials and it’s very lightweight and efficient. Microsoft is going to offer Security Essentials for free for all Windows users so that should help out.

No Repository System for Windows and no, good site for Programs- This much is very true, I kind of wish there was something like that for Microsoft that’s similar to Apple’s site. However, then Microsoft would be viewed as a monopoly because they would probably have to choose which applications are showed on the front page.

No PDF Viewer- Yeah, it doesn’t have a PDF viewer but I believe Microsoft tried to add a PDF viewer with the office suite but Adobe sued Microsoft.

No Zip- It can kind of Zip programs but yeah, it does need a better zip program to go default

No DVD burning Program- Windows 7 can burn DVDs and it can even burn ISOs

No CD to MP3 ripper – Media Player has been able to do that since XP and maybe even before that

No Photo Editor – Vista can rotate images easily and it can fix the images, with Windows 7 you have to get Windows Live Photo Gallery to do that but I think it’s widely known thanks to the commercials, it’s also very good with editing photos.

No basic Codecs (DIVX) – Windows media player 12 comes with DivX as one of it’s codecs it even comes with .mov

Unpolished Software – Windows 7 seems very polished to me, I wouldn’t expect a windows operating system to be able to code in UNIX.

Difficult for Developers – I can’t say much about that, I code with C++ and Visual Basic and it’s not too difficult to work with at all.

Not Tunable for Hardware Makers – Microsoft Tuned one of Dell’s computers and they were able to reduce their boottime by 5 seconds. Obviously Windows works with netbooks so why does it need to be tuned?

Until TextMate came along, one might say, and picked up on it.

Whether I’m simplifying or not, I think I have a point that much of TextMate’s “Emacs features” are also readily available in Vim. (I note that the same is often true of the “Emacs features” that TextMate lacks, such as split windows.) TextMate’s line-wise handling strikes me as particularly vi-ish, though, like copying the current line without a selection. Emacs doesn’t have anything like that out of the box (it prefers to work with sexps); vi was always better at handling single lines than Emacs.

Perhaps the community dwindles when TextMate 2 is delayed, but I don’t think it will matter once it’s released. Large amounts migrated from Emacs and Vim to the first version; if the next is even better, more switching is to be expected.

My editor “timeline” over say the last 10 years goes sorta goes like this (definitely missing a few):

pico ->Ultraedit -> Emacs -> VisualAge Smalltalk -> JBuilder -> Emacs -> VS6 -> Emacs -> VS.NET -> Emacs -> Eclipse -> Textmate -> Emacs -> Vim

Although I usually used Vim for quick editing files in linux, I never really made it my “development environment” full time. Every time I tried in the past, I just kept going back to Emacs.

So what made me stay this time?
-The fuzzyfinder plugin which lets me switch files and buffers quickly is a great Textmate replacement for comamnd – T
-Nerdtree is a good source tree explorer
-A bunch of plugins from Tim Pope like rails.vim, and surround.vim are just too useful for me not to use it.
-The general multi-mode Vi awesomeness.

Textmate’s lack of split windows always caused me to keep looking at other editors.

One thing that Textmate did do though is kick all the other editors in the ass for a bit causing them to add things like snippets etc.

I think the biggest mistake they are making is not releasing Textmate 2. There are a lot of people in the Rails community at least that have switched to Emacs or Vim, and I don’t see them coming back.

The biggest reason though for switching is that I have to switch every once in a while. It’s a running joke around here that I can’t use the tools for too long. We’ll see how long Vim lasts

As for Twitter and risk – Scala is a substantially smaller risk than many new-ish languages. Its runtime is basically just the JVM – there’s very little runtime support needed by Scala. That contrasts with a language implementation like JRuby where a great deal of complicated runtime machinery is need to support Ruby’s dynamic meta-programming features. As an example of what that buys you the FindBugs “lint checker” works well with Scala generated byte code where it’s a hopeless muddle with the more dynamic JVM languages. Still, it would be silly for me to suggest that it’s as safe and conservative as Java or C++.

Only time will tell if Twitter made the right decision. But I do know that they tested the heck out of the Scala based message queue before plugging it into production. I also know they’ve been using it for nearly a year now – all this flap is actually about old news.

In what ways did you find TextMate lacking?

I think there’s a tendency to overemphasize “newer is always better” in the software world. It’s far more important that something stands the test of time, in my opinion …

For my current configuration you can check out: http://github.com/marcjeanson/vimconfig

I had tried viper out way way back and wasn’t impressed. Not sure about vimpluse.el. Not sure I’d use “unbeatable” and Visual Studio in the same sentence, although resharper is pretty good

I used to use emacs with visual studio using some add-on that I can’t remember at the moment. It was OK, but less than optimal.

In the end, I think it’s weird and a bit sad that the best editors (Vi/Emacs) were created in the 70′s and we haven’t figured out a better way yet.

I doesn’t have to be “IDEs vs. Editors”, though. Visual Studio + ViEmu is probably an unbeatable combination. If you’re into Emacs, there’s Viper and vimpulse.el.

Microsoft could just make, release and support C# + runtime on Linux themselves. You know, like Sun does with Java. That would show they really care about multi-platform support.

Back when MS was pushing DCOM, they did the exact same thing. They contracted out the DCOM support on UNIX to a third party developer, and pointed anyone to it if they asked about non-Windows support. It too played catchup and constantly fought for relevance.

Cheers

And what is Microsoft supposed to do about that? Give Mono access to their code before its finished? Update the language specs before they’ve been implimented by Microsoft (not possible, specs change during implimenation)?

C’mon. Mainly, you’re trying to justify your dislike for Microsoft, even after it has enthusiastically supported Mono.

I’ve been using Maven for 3+ years, I don’t do it for any “cool” factor, I do it for easy of builds and baked in dependency management.

The point of the post was, why do corporations insist on each project being structured and built exactly the same way? Why do they care? “So anyone can go in and instantly understand your program” might be an excuse, but it’s too much of a price to pay to have that level of transparency IMHO. Just document how you are building things and where things exist and you’ll be fine.

Especially in this case where Maven was supported by their CI server, but because their Build Manual didn’t have a section on Maven, it didn’t exist to them. Also, there were tonnes of special build scripts and what not attached to their ant scripts, things I could not have known about and things I could not examine because they had my account locked down so tight I couldn’t fart without permission.

I guess I’ve just seen too many Windows-centric projects, and am very skeptical that Windows developers look forward to other platforms when building their applications. When it comes time to port it, if it’s not easy, they give up and ignore Linux (and why shouldn’t they, it has a tiny desktop market share). I think you’re overestimating Mono’s current compatibility greatly. But I wish you luck in your project and would love to hear that you had a swimmingly easy time in your port.

@Roy

Lost it to Mono? I don’t have any Mono applications (large, non-core-system applications) installed, yet I have a plethora of C++ and Java desktop applications. This is especially true when you look at commercial applications, who are more pragmatic and conservative when it comes to choosing languages and libraries. I’m not biased against Mono apps though. If I find one that I like, I won’t think twice about apt-get’ing the required runtime.

Alex, your main issue that Mono will always have to play catch-up to the MS releases is valid, but from a practical standpoint I don’t think it is detrimental. I recently went to one of the Visual Studio/SQL/Windows 2008 launch events and trainings to see the latest and greatest. In one of the seminars they did a hand raising asking how many people had migrated to .NET 3.0, how many were using WCF/WPF, and a bunch of other year-old and two-year old language improvements. The truth is, maybe 10% of the hands went up. Adoption of the new stuff and especially incorporating it in larger projects with existing coding standards takes a long time — most people weren’t beyond .NET 2.0 framework. I believe Mono is already fully compatible at 3.0, or further.

I and other developers I work with are already evaluating if we can move some projects to Mono. Hell, I even attended SoCal Linux Expo last week for the first time because I wanted to see what else was out there (and learned a ton!).

Also, I think the analysis about Java is fair, almost all my university schooling was in Java, and C# borrows from it in the areas I liked, and improves some of the areas I didn’t like about Java. The languages are so similar though, in my opinion it would be really hard for Microsoft to go after Mono on C# considering how much Microsoft “borrowed” from Java. Same with Moonlight, especially considering how MS folks stayed up late night before the Obama inauguration to help get Moonlight capable of streaming the video, can they turn around and sue Novell now? Anyways, it’s a good debate. Viva Mono & viva Miguel!

I just guess ultimately we’ll see in a year or two and what percent of C# apps or silverlight apps are actually usable under Mono/Linux.

I don’t attack Mr de Icaza personally (as I don’t know him), and I highly respect the work he’s done for the Linux Desktop. He is one of the big leaders in the community. Regardless, that doesn’t make every decision he makes perfect. In my personal opinion, stressing C, ignoring C++ and going the C# route was/is a mistake, as outlined in my post. Nor do I agree with his defense of MS in his public statements in the podcast. These are my personal opinions and I’m entitled to them.

I don’t criticize the Mono code itself, which I’m sure is of high quality. Again, I’m just saying they’re always going to be chasing ever changing Microsoft standards, and always playing catchup. That is all.

Please don’t put words in my mouth.

I mentioned that I was using Debian *as context*, a way of indicating that I have no particular positive feeling about Microsoft products. I realise now that this was a concept too far.

I’m not sure what MS has to do with world issues such as hunger or pharmaceuticals, as that’s beyond the scope of my post. MS is still a monopoly, even if software operating systems aren’t a basic need for human life. These are just facts, I don’t see how this can be perceived as religious. I need to state these facts (even if we’ve all heard them a hundred times) to frame my views on Mono.

I never said Mono isn’t free software. I’m just saying they’ll never attain compatibility with the latest MS versions. It’ll always be in catch up mode. There will always be pain with it. I’m just hoping that people are aware of this going in – that is all.

I’m writing this from my Ubuntu desktop. I could have just as easily wrote this from my Windows XP development box or one of my XP laptops, though. I’m not sure what this proves

I’m writing from my Debian desktop which doesn’t contain any mono libraries or applications, a choice made only on performance grounds and for the same reasons that I don’t use any free software Java based apps. Mono is free software by any standard and people making snide and sly remarks about Miguel de Icaza doesn’t change that. Stockholm Syndrome indeed.

“The question for the community will be: why should they deal with a group whose influence takes second place to Microsoft’s overall objectives of protecting patents and selling Windows? This at a time when Microsoft has been reaching out to open-source developers to help plug-ins and models that will extend Visual Studio 2010 and its new Oslo modeling platform.”

from El Reg: http://www.theregister.co.uk/2009/02/26/microsoft_tomtom/

And put this in it:
escape ^Zz