Aldo Cortesi

Ew! :) - Pistos

"The fact that Perl is not consistently parseable is hilarious, but not the core issue. There is more to language size than the size of the parser. Perl has consciously tried a design philosophy that gave up a focused language design in favor of "there is more than one way to do it". This was a fine experiment to try - it's not obvious before the fact whether this approach this would work. Even die-hard Perl bigots have to realise, though, that the tide of popular opinion is beginning to turn against them, and that developer mindshare is moving towards Ruby, Python and maybe even languages like Haskell and Erlang. Now, it may be that this is because the universe is deeply unfair, and that this sorry state of affairs can be corrected by making pissy comments on other people's blogs. Then again, it might just be that there's something more solid at play." - Aldo Cortesi

"Does EVE have a scripting interface? I looked at it a while ago, but it didn't look like it had any in-game scriptability - WoW's Lua interface gives a clever programmer a great toe-hold for things like trading applications." - Aldo Cortesi

"I'm not convinced by any of this. I'd be surprised if Perl projects are generally more test driven than Python projects. The "cultural virtues" angle is too fuzzy, and it's not clear that there's any difference between Perl and the general Open Source community. For what it's worth, my comment about Perl in the article was off-hand speculation, and the venomousness of some of the Perl commentators here and in private mail has taken me aback. I do think that the Perl community feels that it is on the back foot at the moment - the wide perception is that the Perl approach was a noble experiment, but that it has lost in the language wars. If you'd like to make a case that the commit rate in Perl projects drops off because Perl projects tend to approach perfection more rapidly than those in other languages, be my guest. I think it's a very silly argument, but I've shared the data exactly so that other people could take an independent look at things." - Aldo Cortesi

"At the risk of inflaming more Perl programmers to come and manfully defend their language on my blog, I think there's a reason why Python has a nice BNF grammar, and Perl has 5600 lines of ad-hoc parsing code: http://www.perlmonks.org/..." - Aldo Cortesi

"At the risk of inflaming more Perl programmers to come and manfully defend their language on my blog, I think there's a reason why Python has a nice BNF grammar, and Perl has 5600 lines of ad-hoc parsing code: http://www.perlmonks.org/..." - Aldo Cortesi

"These are the Haskell projects that passed all the rounds of elimination: hack gitit scion nemesis polyeuler turbinado dyre mpdfs hPage bert yuuko loggerbot restful compleat The "polyeuler" seems to be a set of solutions to the Euler project in various languages, that just happens to have more Haskell code. attempt loli monadlib Hakyll" - Aldo Cortesi

"No particular reason. 300 is a nice round number. ;)" - Aldo Cortesi

"Nope, this can't explain the Perl data at all - remember that I'm using information for Perl projects that _already_ use github. Whether or not github is popular among writers of Perl is irrelevant, unless you're suggesting that the poor saps are gradually worn down into reducing their commit rates by the grinding horror of using github, but never bother to move their projects elsewhere." - Aldo Cortesi

"You want me to respond seriously to the claim that Perl projects see a sharp decline in commits over time because they somehow become "complete"? Unless you're Donald Knuth, that's a hilariously ridiculous claim. There's nothing that makes Perl special in this respect." - Aldo Cortesi

"Believe it or not, matplotlib, with much coaxing, coddling and wheedling. I implemented the curve fitting and smoothing algorithms from scratch." - Aldo Cortesi

"Well, I don't think my reaction is _that_ mysterious. ;) C is a small, compact language, and it's used successfully by some of the very biggest software projects in the world - the Linux kernel, for instance, is probably the biggest most active software project of all. C scales astonishingly well. Its faults - and I'm not claiming it doesn't have them - are elsewhere. Perl, on the other hand, is a language so large that no-one knows it in its entirety, so everyone writes code using their own subset. I could well believe that as projects grow and gain committers it becomes harder and harder to keep things together. For what it's worth, I've worked with and audited some really large Perl projects, so I'm not speaking entirely without the benefit of (bitter) experience." - Aldo Cortesi

"No, I think this is a great observation. That graph _does_ include all lines - the information is essentially just what git reports for the patch statistics. So I think you're probably right." - Aldo Cortesi