Front page | perl.perl5.porters |
Postings from July 2010
Nicholas Clark <nick@ccl4.org> writes: > On Thu, Jul 15, 2010 at 04:46:28PM +0200, Steffen Schwigon wrote: >> - Why do companies invest effort into other languages? >> >> - Why is it more difficult for a company's employee to contribute to >> Perl than to, say Java? > > Do they, and is it? > > Did anyone other than Sun invest in their Java VM and bundled > technologies? Are their corporate contributors to other Java VMs? IBM once had a VM, AFAIR, not sure. However, I admit that Java might be a Sun driven special-case; strong marketing + lucky point in space-time continuum. > I don't see anyone pointing out good examples of other firms > contributing to other languages, other than Google on Unladen > Swallow (has it stalled?) Facebook on Hip Hop (can Media Wiki run > on it yet?) and the Intel Haskell work mentioned below. But that's the scale I am talking about; major languages, major companies. >> - Why is Java a business-case for so many but not Perl? > > At a guess > > a: Because it's designed to scale to a lot of mediocre programmers They probably also got some clever ones. :-) > b: Universities etc have been persuaded to mass produce Java > programmers Indeed. But by whom? IMHO it's already a by-product of having companies involved. That's why I'm riding that point. > c: There's a certification scam, er scheme. It sounds negative but at least that seems to help. I even remember discussions about that for Perl. I think it's something that happens automatically once there's a “corporate” momentum. But as I already admitted, maybe Java is special. >> - Why does Intel care for Haskell parallelism but not for Perl? > > That is a good question, and the best answer I can suggest is > > Intel can't make their CPUs faster, so they make them with more > cores. Thus parallelism is forced upon them. [...] > > Implementing parallelism is much easier in a language where (most? all?) > variables are read only rather than read write, such as functional languages. > Of the functional languages, Haskell is the closest to going mainstream. > So it's a good target. Better than Perl, Python, PHP, Ruby or Lua Plausible. >> - What is Perl's business case? >> (Its hidden omnipresence isn't enough, obviously.) >> >> I don't think it's about hacking the Core. Other compilers are also >> non-trivial. And there is lot of work possible before hacking the >> core. Analyzing, debugging, benchmarking, sponsoring hardware, CPAN >> interfaces to proprietary products, XS wrappers. > > What do you mean by "Perl's business case"? > > Pitching to whom? Companies who are potentially going to do what? Like the intel thing above. Big scale, big sale. Via whatever indirection, and leveraging on the language level provides a huge indirect multiplier to all applications of that language. And it's the same like my other examples around Google/Python, Facebook/PHP. Their involvement makes them more competitive. Does helping Perl make someone more competitive or sell more? A big spam filter service provider who tunes Perl to speed up SpamAssassin would be such a case. Or hardware OEMs selling servers to such spam filter companies. Or companies providing major components to such OEMs. Unfortunately I'm running out of such examples. SpamAssassin is my strongest one. I already asked around for more in several places, SpamAssassin, BioPerl, TPF, EPO, etc. All in all, I think having *some* kind of corporate momentum would help Perl. That's why I still support the general idea of Gabor. Kind regards, Steffen -- Steffen Schwigon <ss5@renormalist.net> Dresden Perl Mongers <http://dresden-pm.org/>Thread Previous | Thread Next