LWN.net Logo

Videos from the 2010 LLVM Developers' Meeting

Videos from the 2010 LLVM Developers' Meeting

Posted Dec 22, 2010 2:20 UTC (Wed) by HelloWorld (guest, #56129)
In reply to: Videos from the 2010 LLVM Developers' Meeting by xtifr
Parent article: Videos from the 2010 LLVM Developers' Meeting

There's nothing wrong with having different languages for different purposes, but every time somebody writes a library in a language I don't use, that means I can't use that library. I also don't think that all those new dynamically typed scripting languages have anything meaningful to offer over Lisp or Smalltalk, both of which have been around for much longer than Python, Perl, Ruby and their ilk.


(Log in to post comments)

Videos from the 2010 LLVM Developers' Meeting

Posted Dec 22, 2010 8:57 UTC (Wed) by butlerm (subscriber, #13312) [Link]

Clearly LISP and Smalltalk have a problem, or they would have taken over the world decades ago. The difficulty with LISP is obvious enough, with Smalltalk not so much.

Videos from the 2010 LLVM Developers' Meeting

Posted Dec 22, 2010 11:44 UTC (Wed) by sorpigal (subscriber, #36106) [Link]

The problem they have, the problem all languages have, is that language choice isn't always a purely rational.

Videos from the 2010 LLVM Developers' Meeting

Posted Dec 24, 2010 6:43 UTC (Fri) by jmalcolm (guest, #8876) [Link]

In my mind, the logic here is flawed.

First, it is not a given that the person that wrote the library you cannot use would have written one that you can even if there was only one programming language. Something inspired them to write that library and your "good enough" default language choice may not have done it.

Second, competition results in tremendous redundancy I agree. However, it also seems to result in greater capability and greater accessibility as well. Which system produces better products: communism or capitalism? I think that history shows that capitalism utilizes resources far more effectively. We should always cheer greater competition.

Certainly it is annoying to have more choice than you want, especially if it means you have to make a bet on where to invest yourself (in terms of time and expertise). I do not agree though that we are somehow being held back by all these people writing great software in languages we have not yet chosen to use.

Quite the opposite.

Videos from the 2010 LLVM Developers' Meeting

Posted Dec 25, 2010 3:24 UTC (Sat) by HelloWorld (guest, #56129) [Link]

> First, it is not a given that the person that wrote the library you cannot use would have written one that you can even if there was only one programming language.
I think that people write libraries because they need them to get something done, not because the language inspired them to. If that is so, then the language doesn't matter and they would just have picked another one, as the thing they need to do doesn't magically disappear.

> Second, competition results in tremendous redundancy I agree. However, it also seems to result in greater capability and greater accessibility as well. Which system produces better products: communism or capitalism? I think that history shows that capitalism utilizes resources far more effectively. We should always cheer greater competition.
The computing field is fundamentally different from any other market, as compatibility and interoperability are both more important and more difficult to achieve than elsewhere. Thus, very often choices are made not because of the merit of the product itself, but because of compatiblity constraints. Therefore, competition doesn't work as well as in other markets. Actually, competition often led to failure. UNIX vendors fought amongst themselves, allowing the rise of Windows NT. And now, Linux Distros are competing against each other, and yet no distro has a significant desktop market share.

Videos from the 2010 LLVM Developers' Meeting

Posted Dec 26, 2010 0:20 UTC (Sun) by vonbrand (subscriber, #4458) [Link]

I think you are confusing a coincidence with cause and effect when talking about Linux distributions and "desktop dominance". There is only one MacOS, and it hasn't got dominance either...

Never forget that open source works because there are lots of people trying different approaches simultaneously, and in the end each user picks that what turns out best for them. If that leads to a overwhelming mayority picking the same, congratulations! If not, nobody is (much) worse off in the end.

Videos from the 2010 LLVM Developers' Meeting

Posted Dec 26, 2010 14:09 UTC (Sun) by HelloWorld (guest, #56129) [Link]

I think you are confusing a coincidence with cause and effect when talking about Linux distributions and "desktop dominance". There is only one MacOS, and it hasn't got dominance either...
Lack of fragmentation is a necessary, not a sufficient condition. And the single most successful desktop operating system is Windows, which is not fragmented.
If not, nobody is (much) worse off in the end.
Except that I am worse off. I can't use packages that were built for, say, Fedora, because they don't work on my debian system. And I can't rely on documentation for other distros, because it doesn't apply to the distro I'm using. And independent software vendors have a hard time because it's hard to ship software that will run on all distros while integrating with the package manager etc. I don't think that the benefits of having dozens of distros packaging the same software outweigh these problems.

I'd be willing to make an exception for special-purpose distros like OpenWrt or long-term support like RHEL, but not for the dozens of useless desktop distros like pardus or whatever.

Too many languages

Posted Dec 25, 2010 15:04 UTC (Sat) by Kluge (guest, #2881) [Link]

Maybe it's not too many languages, but too many runtimes. Systems like Parrot or the CLR seem designed to allow mixing of languages.

But, like most, I haven't used them, so I can't say how useful they are...

Copyright © 2013, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds