the current german computer magazine iX has a report about mobile.de having migrated silently from Perl to Java in the past two years. The essence of the report in short:

All in all, for a person who knows (e. g. me *g*), it is evident, that the problem was not so much Perl as technology, but management constraints. Nevertheless, for a Perl advocate this is Waterloo, as now every manager in Germany will point to this case and we will probably see more large scale projects dying.

Not because Perl is bad, but because you simply do not get enough people at reasonable cost that would master Perl for large-scale projects. Either they are incompetent, or the competent ones are bound elsewhere, or they are competent and available, but pure prima donnas and do not play well in teams, or ... ;-)

But that's another story.

Bye
 PetaMem
    All Perl:   MT, NLP, NLU

Replies are listed 'Best First'.
Re: Report: Migrating from perl to Java
by shmem (Chancellor) on Mar 08, 2008 at 13:58 UTC
    system being alegedly x times faster, taking x times less ressources etc. etc.

    I guess they would have achieved the same overall effect migrating from perl to perl, i.e. refactoring for 2 years... which endeavour would probably have reduced that 550 kLOC, who knows - hard to tell not knowing what the perl code looked like.

    So it is hard impossible to tell (and misleading doing so) how great the net benefit of porting to java has been in terms of performance and resources. The biggest benefit clearly is having a unified code base and a single programming language.

    update: s/hard/impssible/

    --shmem

    _($_=" "x(1<<5)."?\n".q·/)Oo.  G°\        /
                                  /\_¯/(q    /
    ----------------------------  \__(m.====·.(_("always off the crowd"))."·
    ");sub _{s./.($e="'Itrs `mnsgdq Gdbj O`qkdq")=~y/"-y/#-z/;$e.e && print}

      If you are only interested in the small pic...

Re: Report: Migrating from perl to Java
by perrin (Chancellor) on Mar 08, 2008 at 13:38 UTC
    I would expect eBay to want to move sites they purchase to Java, since that is now their platform. (They used to use Windows DLLs in C++.) It's interesting to note that eBay is nearly the only large site that has gotten Java to work for high volumes. The problem is exactly what you describe in your last paragraph -- not so much a Java issue as one of finding talented people.
      The cost is another main issue, if not bigger. To build apps on a platform costs much more effort and time than to just build it.
Re: Report: Migrating from perl to Java
by jepri (Parson) on Mar 08, 2008 at 13:34 UTC
    I've been reading articles about how perl is dying ever since I started learning it. They used to distress me, but now I see them as a sign of perl's health. If I don't see a 'perl is dying post' for a few months I get edgy.

    Thanks for topping me up! That should get me through the next few months.

    I take heart from the fact that the large website was originally written in Perl. And I'm kinda impressed that they managed to complete a Java project on time. I thought Java was supposed to prevent that sort of thing happening.

    ___________ Jeremy_______ Bots of the disorder.

Re: Report: Migrating from perl to Java
by Your Mother (Archbishop) on Mar 08, 2008 at 20:30 UTC

    Just for fun. I suspect this is an underestimate and does not include the hardware upgrades that seem to come with Java. A lot of these "success stories" look like pork and Chinese firedrills when viewed from inside a company instead of from reading the PR releases.

    my $full_time_devs = 20; # Guess my $annual_work_hours = 1800; # Guess my $hourly_us = 40; # *Total* guess my $years = 2; printf("\$%s\n", commify( $full_time_devs * $hourly_us * $annual_work_hours * $years ) ); sub commify { my $text = reverse $_[0]; $text =~ s/(\d\d\d)(?=\d)(?!\d*\.)/$1,/g; return scalar reverse $text; } __END__ $2,880,000
Re: Report: Migrating from perl to Java
by PetaMem (Priest) on Mar 09, 2008 at 16:45 UTC

    Thanks for your comments and insights. Just to clarify:

    • I too don't think this means Perl is dying. Also this wasn't a PR news, but from a reporter of the magazine itself (neither eBay nor mobile.de wrote it).
    • My first thought was also, that refactoring perl for two years would bring probably similar results BUT and this is what made me think: only if you get the right people.

    And that is what my last paragraph was about. Most of the time, you don't get the right people for large-scale Perl projects.

    Nevertheless - one big supporter of the annual German Perl Workshops is gone. Let's see if this gap can be filled in the future.

    Bye
     PetaMem
        All Perl:   MT, NLP, NLU

A reply falls below the community's threshold of quality. You may see it by logging in.