[OT] Re: Perl Doesn't Suck
schwern == schwern [EMAIL PROTECTED] writes: schwern Sun doesn't give out its JDK source code freely, they schwern have all sorts of restrictions. If I wanted to port the schwern JDK I can do it, but I need special permission from Sun schwern to distribute it. This Sucks. Especially considering schwern that you could go through a whole lot of work and then schwern have Sun deny you. Speaking as someone with feet firmly in both camps (I'm a Blackdown member and the Debian maintainer for the jdk and some of the largest perl modules in that distribution), IMNSHO the fatal assumption made by millions of people is that Java is OpenSource just because it's a *language*. It isn't. Get over that expectation and life becomes simpler. Just because there's a Fortran or a COBOL standard doesn't make it any easier to write compilers and run-time support from scratch for those languages, regardless of platform. And, no, no one will help support you in this endeavour. That's where things like Kaffe and Japhar find themselves. Perl's great blessing is also it's great curse; there's a single implementation and that *implementation* happens to be OpenSource. Try writing a second Perl implementation from scratch. Were something dreadful to happen to Larry and his estate chose to change the licensing terms of the current *implementation*, where would Perl6 go? Or even Perl5 for that matter? -- Stephen Farcical aquatic ceremonies are no basis for a system of government!
Re: Perl Doesn't Suck
On Fri, Jun 29, 2001 at 06:59:07PM -0400, Michael G Schwern wrote: As a Solaris gal, this might all seem perfectly sensible. Myself, I'm doubly screwed running Linux (which is not Windows or Solaris) on a PowerPC (which is not Intel or Sparc). Sun might have perfectly valid business reasons for concentrating on Solaris and Windows, but in the end it still means I'm screwed. Blech. Sun didn't force you to run either Linux, or to use a PPC platform. It's your choice, and blaming Sun to screw you doesn't make much sense. Abigail
Re: Perl Doesn't Suck
On Fri, Jun 29, 2001 at 05:29:53PM -0500, Elaine -HFB- Ashton wrote: Not all OS, though most, have Perl in the base install and those that do even have problems. Config.pm has issues on HP and Sun, RedHat has spotty RPMs that occsaionally go awry. That's their fault. Find a better distribution. *Then again, even for big things like Tk or LWP, I've found perl's *installation process to be at least more consistent. perl Makefile.PL * make test make install will just about do it. By contract, I've *had to install the JMF under Linux. HOoboy, that was a mess. All *sorts of little hard-coded paths and config options to tweak and *jigger, copying things around by hand. Unpleasent. I'd imagine it *would be easier to do on Solaris, but that doesn't help me. Well, call your Linux support and ask for them to make it better/easier to install Java on Linux. Sun isn't going to do it for the other platforms I would imagine. Well, see there's the trick. Sun (and thus Java) cares about Solaris and Windows. Everyone else has to fend for themselves. To make matters worse, its very difficult to get your hands on the JVM source code, so its not like I can even fix it. With Perl, I can at least fix it myself and roll it back into the core. Its the old Open vs Closed Source argument. As a Solaris gal, this might all seem perfectly sensible. Myself, I'm doubly screwed running Linux (which is not Windows or Solaris) on a PowerPC (which is not Intel or Sparc). Sun might have perfectly valid business reasons for concentrating on Solaris and Windows, but in the end it still means I'm screwed. With Perl, if you're running some fucked up OS on some fucked up hardware we'll at least give it our best shot to get it working. If you figure out how to get it working, we'll do what we can to roll it into the core (mod a little p5p squabbling). The one time I did have an issue with a Java package I had an engineer issue a patch within the day...and he was awakened at 4am. Did Sun do this out of the goodness of their heart or did you have commercial support? Given that Sun charges $200 per incident for a two day response time ($1600 for four hour reponse) this is not something most people can afford. Expensive, commercial support, while it is nice for those who can afford it, is not a general solution. *But yes, module installation can be made easier. We're working on it. No, module installation isn't hard and wasn't what I was driving at. Enterprise wide deployment. Make it easy to install and deploy perl across 5k boxes in a farm and you'll have a winner. :) With Debian I just make dpkgs (or use the Debian ones), stick them into a local repository and sync normally from there. Most OS's have a similar packaging system. Otherwise, you can use CPAN::Site to set up a local CPAN repository which overlays your own. Or you can set up your own local CPAN mirror of approved modules. After that its all cron jobs. Making it easier to do those last two is a good way to go. That an Ingy's NAPC idea. Like I said, we're working on it. -- Michael G. Schwern [EMAIL PROTECTED]http://www.pobox.com/~schwern/ Perl6 Quality Assurance [EMAIL PROTECTED] Kwalitee Is Job One Only mindless violence can raise my spirits now!
Re: Perl Doesn't Suck
On Fri, Jun 29, 2001 at 05:20:40PM -0400, [EMAIL PROTECTED] wrote: There's the trick, Solaris is Sun's Blessed Platform. As a Linux/PowerPC user, I know how Ziggy feels. I'm almost totally ignored by Sun and I'd imagine I'd have just as much trouble getting it working as he did. This is the issue in a nutshell. Let's not mix business issues with technical ones. Let's not mix cluster management with simple end-user installation. Let's not mix businesses losing millions by the microsecond with the guy who just wants his little website to just *work*. [*] Personally, I don't really care what my support options are at 4am, who I can pay to rouse out of bed, or whose pager I can pay to page. I want technology that works, not technology with a glitzy sales presentation and more hype than the Beatles. I don't want technology that's unnaturally bound to customer support to make it work. I want something that works, as advertised, and works reliably enough for me to adopt it (perhaps after making a few patches). Plenty of companies find lots of work playing in Sun's (or Oracle's or Microsoft's ) view of the world. That does not define the world of computing. Not by a longshot. Let's not forget that the mainframe never died, even after 20 years predicting it's imminent demise. Remember too that the proprietary, commercially supported client/server programming languages focused on mainstream platforms like Win3.1+Sybase are now a footnote to history. This is especially sad with Java, which promised to be write-once run-anywhere, and continues to fail to deliver on that promise. This is especially nice with Perl, which has delivered on write-once run-anywhere without promising it, and has done so for more than a decade. Doubly so when Perl's single-implementation standard gives predictable results pretty much everywhere, while Java is subject to multiple JVMs which can be differently buggy. But yes, module installation can be made easier. We're working on it. There are always improvements to be made, even with Perl. But where we are today, Perl doesn't suck. Z. *: What good is that multi-billion dollar business doing? Who knows. What good is that little guy doing? Who knows, but it might be the next version of the OED, a realtime snapshot of the company's PL, or satellite data that reaches you quicker.
Re: Perl Doesn't Suck
Adam Turoff [[EMAIL PROTECTED]] quoth: * *This is the issue in a nutshell. Let's not mix business issues *with technical ones. Let's not mix cluster management with simple *end-user installation. Let's not mix businesses losing millions *by the microsecond with the guy who just wants his little website *to just *work*. [*] P5P was created becausePerl sucked to get it installed on many platforms. If Java sucks to install on some boutique/niche platforms it could mean that a) noone has told them about the issues or b) noone in the FreeBSD/Linux world has taken it upon themselves to care enough to love it and see it through to make it work. What's the 'little guy' using Java for anyway since Open Source supposedly should have revolutionised the choices for them by now. Where would Perl on VMS be without Peter Prymmer? etc? The issue behind platform selection and support is that since Sun is a company out to make money, it's going to choose the ones that will reach the most interested paying customers. Some would even say that makes a certain amount of business sense, especially in an economy where tech profits are drying up. Java has no platform angels in the corporation or out in the field most likely. It doesn't have to mean it sucks rather it is a function of it's market. *Personally, I don't really care what my support options are at 4am, *who I can pay to rouse out of bed, or whose pager I can pay to *page. I want technology that works, not technology with a glitzy *sales presentation and more hype than the Beatles. Probably because you aren't a Sysadmin called to fix something a developer broke who doesn't wear a pager and doesn't answer the phone at 4am so you are left with fixing the cryptic code yourself or calling support. It's not about hype, it's about time. Linux used to be a real bitch to install on non-supported hardware too, especially the esoteric and the new. You learned to buy supported hardware or live with waiting for a driver. The same could be said for the FreeBSD and even Solaris on Intel. It takes time. *There are always improvements to be made, even with Perl. But where *we are today, Perl doesn't suck. For you. It probably does suck for some people which is why we should be glad that there are so many platforms and other languages to choose from. If something sucks, use something else. After losing yet another platform, Alpha, this week I would think that we should cherish every bit of diversity in computing we can get. e.
Re: Perl Doesn't Suck
On Sat, Jun 30, 2001 at 11:57:42AM -0500, Elaine -HFB- Ashton wrote: If Java sucks to install on some boutique/niche platforms it could mean that a) noone has told them about the issues I can't even conceive they're not accutely aware. b) noone in the FreeBSD/Linux world has taken it upon themselves to care enough to love it and see it through to make it work. I think its c) WE HAVE NO SOURCE TO WORK FROM! Sun doesn't give out its JDK source code freely, they have all sorts of restrictions. If I wanted to port the JDK I can do it, but I need special permission from Sun to distribute it. This Sucks. Especially considering that you could go through a whole lot of work and then have Sun deny you. Some are trying from scratch, a truly masive undertaking given that the current JDK is 20 megs *compressed*. Hungry Hackers has Japhar which is a 1.1.5 JDK and there are others, but it all lags behind. If Sun really wanted to help porters, they could release the JDK source code Freely. They won't. Do not blame the Open Source guys for lack of effort. Where would Perl on VMS be without Peter Prymmer? etc? White Knights are necessary, but Peter Prymmer did not have to rebuild perl from scratch. Being an OS Knight isn't an easy job, but we don't throw roadblocks in their way. Some of us have been bending over backwards to help. Guys like Hungry Hackers have 0 help from Sun. All they've got is a spec. The issue behind platform selection and support is that since Sun is a company out to make money snip That's all prefectly valid. It still means Java doesn't work well on my machine. Java has no platform angels in the corporation or out in the field most likely. Hungry Hackers and Blackdown are two off the top of my head. Blackdown is in a special position in that Sun has granted them permission to use the JDK source code, but nobody else can see it. If I don't have the source code, how am I supposed to fix anything? I can patch something and send it off to Sun, sure, but let's think about a full platform port... Let's say Peter Prymmer decides he wants to port the JDK to VMS (God Help Us ALL). What's his course of action. 0) Consult a psychiatrist. 1) Joins the Sun Developer Community 2) Download a copy of the JDK 3..100) Goes through a massive amount of work to get it working under VMS 101) Trumphantly sends the patches off to Sun. 102a) Sun sets Peter up with a Blackdown-like deal allowing him to distribute his port. **OR** 102b) Sun thanks Peter for his time but isn't going to trouble itself with VMS. **OR** 102c) Sun thanks Peter for his time, but doesn't feel the port is good enough to be blessed. Its that last bit of fuzziness at the end. You could go through that work and still be denied at the end. At least, that's how I'm reading their license. The other problem is steps 3 to 100 must be done in a relative vacuum. I don't believe he'd be allowed to redistribute his patches, so no vmsjava mailing list or community of developers to work with. Of course, he could ask permission from Sun beforehand, but again, this all relies on the good will of Sun (and them trusting Peter, who could be some schmo for all they know). Perhaps more importantly than all this, the spontaneity is lost. I know of many, many porting projects which started out with one guy saying Gee, I wish X worked on my OS and in a flurry of wild patching something is made to work in a few hours/days. Not perfect, but enough to get the snowball rolling. Finally, if the core developers (ie. Sun engineers) are unconcerned with your OS, they're going to keep cranking out platform specific code which you're going to have to keep patching over and over again. At least with perl we've got things like Windows and VMS in the back of our heads when making core patches (and if we don't we'll have Peter BEATING the backs of our heads ;) Thinks either work out of the box or are caught very early on. None of this ports lagging behind a few versions while they fix everything up. Anyhow, this is just the old Open vs Closed source debates rehashed. Except s/Closed/Community License/ -- Michael G Schwern [EMAIL PROTECTED] http://www.pobox.com/~schwern/ Perl6 Quality Assurance [EMAIL PROTECTED] Kwalitee Is Job One
Re: [OT] Re: Perl Doesn't Suck
On Sat, Jun 30, 2001 at 01:49:45PM -0700, Stephen Zander wrote: Perl's great blessing is also it's great curse; there's a single implementation and that *implementation* happens to be OpenSource. Try writing a second Perl implementation from scratch. Fortunately, we don't have to. :) Perl 6 should have a more fathomable design, or at least better documented. I don't think it'll ever reach the sort of standards that things like Java and C++ have. Then again, have you LOOKED at the ANSI C++ standard? ;) Were something dreadful to happen to Larry and his estate chose to change the licensing terms of the current *implementation* In that Highly Unlikely event, we can simply fork off the source code from the point where the license changed. License changes are not retroactive. The only restriction is we couldn't call it 'perl' under the AL. I don't think such a thing has ever happened to a major Open Source project, this doesn't worry me. The license squabbling between the various BSD's is similar, but they're all still Open. where would Perl6 go? Same place its going right now, as there's no code. :) -- Michael G Schwern [EMAIL PROTECTED] http://www.pobox.com/~schwern/ Perl6 Quality Assurance [EMAIL PROTECTED] Kwalitee Is Job One
Re: [OT] Re: Perl Doesn't Suck
Stephen Zander [[EMAIL PROTECTED]] quoth: * *Speaking as someone with feet firmly in both camps (I'm a Blackdown *member and the Debian maintainer for the jdk and some of the largest *perl modules in that distribution), IMNSHO the fatal assumption made *by millions of people is that Java is OpenSource just because it's a **language*. It isn't. Get over that expectation and life becomes *simpler. Well, I think a lot of people don't consider the OS v CS when choosing a technology and, honestly, it's not the point I was trying to make. I see too much griping about how anything that isn't open source sucks, etc. while all around us there are fewer and fewer choices to make. I don't care so much about licensing as I have my own personal choices that aren't shared by everyone nor do I expect them to be so or consider others who don't to be stupid. But I am very concerned that as time goes by that diversity in platforms and in software will cease to exist. George Carlin had a interesting observation that our lives are filled with meaningless choices at the grocery store, 165 breakfast cereals, over 100 different kinds of mustard, and that while we have these meaningless choices we have little diversity in important choices; 2 poltical parties, 2 or 3 major media companies, 2 home computing platforms, etc. I have this same feeling when looking at computing these days. Saying that Microsoft Sucks has not had much effect on its market share or in deterring people from using it. I wouldn't expect much more of an effect by saying Java or anything else sucks. Obviously, there needs to be a better method of making open source technologies more attractive to more people. e.
Re: [OT] Re: Perl Doesn't Suck
All, of course, imho: Were something dreadful to happen to Larry and his estate chose to change the licensing terms of the current *implementation* Well they can only do that to a copy of their own, not existing copies. While the law isn't clear on a lot of nuances related to more complex open source licenses, perl's is clear enough that it is implausible that A) a non-corrupt modern judicial system is going to interpret it in a way that allows license changes to apply retroactively to existing copies and B) they can then enforce this in a manner that truly stops a fork. (And guess which version the book authors and gurus would follow? Consider Borland's Interbase, which had a far more restrictive license, not even remotely as open as perl. Borland mucked about, so the community forked. All the gurus immediately backed the fork and the new direction was established in less than a week. Now Borland seeks to merge its fork (the original set of sources, since further developed a little) back with the community fork. This will happen, and the pragmatic strength of freedom over self-serving manipulation will have asserted itself.) Try writing a second Perl implementation from scratch. Indeed, Perl 6 is, afaik, a rewrite, from scratch. All you need is a community willing and able to do it. Perl's great blessing is also it's great curse; there's a single implementation and that *implementation* happens to be OpenSource. If anyone concludes that a particular language with a clear spec that stands free of any given implementation is a lesser strategic risk than perl, then I think they should adopt that other language. I very much doubt they will have much success convincing those working on perl that a commitment to a spec that stands free of the current implementation would be a good use of their time, because the license is sufficiently free, and a stand alone spec that kept up with a continuously evolving and pragmatically driven thing like perl would take an inordinate amount of effort.
Re: Perl Doesn't Suck
On Fri, Jun 29, 2001 at 01:18:07PM -0500, Elaine -HFB- Ashton wrote: Adam Turoff [[EMAIL PROTECTED]] quoth: * *Nevertheless, a degenerate case for installing Perl never requires *transfers or temporary disk space measured in quarter gigabytes. Sure it can. Allow me to clarify: a degenerate case for installing a *single* *specific* version of Perl never requires transfers or temporary disk space measured in quarter gigabytes. Mirroring all of CPAN is not a pre-requisite for installing a version of Perl, multiple versions of Perl, etc. *Worst-case-to-worst-case, Perl doesn't suck, and it's doing much *better than Java. I wonder which is easier to support post-install. Perl can suck and often does for the newcomer who, when faced with trying to wade through all the XML modules on CPAN trying to figure out which one is which for what purpose, can be quite frustrated with getting things to work as advertised. That's not the issue here. Chris and I are talking about the case where a user finds a piece of software requiring (Perl|Java) and need to install the language distribution as well as the software requiring that distribution (or, a simpler case of installing a specific version of a language distribution). Wading through CPAN is not an issue here. Everything can suck given enough scrutiny. Surely. And if you want to believe that Perl sucks, please do. While installing Java this week, I stumbled across the installation experience and supported platforms aspects of a language and found that Perl is doing remarkably well, even if no one takes the time to say so. I chronicled my experiences in detail to show how bad it can be, and to highlight how remarkably well-engineered and supported Perl is by comparison. Z.
Re: Perl Doesn't Suck
Having gone through much the same pain a couple of weeks ago (although I just broke down and installed the linux-jdk-1.3.1 port after Sun's web site told me to come back later), I eagerly await a pure-Perl replacement for FOP (http://xml.apache.org/fop/index.html)). -- Christopher Masto Senior Network Monkey NetMonger Communications [EMAIL PROTECTED][EMAIL PROTECTED]http://www.netmonger.net Free yourself, free your machine, free the daemon -- http://www.freebsd.org/
Re: Perl Doesn't Suck
On Fri, Jun 29, 2001 at 06:51:33PM -0500, Elaine -HFB- Ashton wrote: *That's their fault. Find a better distribution. There are a lot of Solaris 8 users out there and to have a broken OEM Perl is not optimal. That response would not be well received. If insert vendor name here distributes a broken Perl with insert OS name here, what are we to do about it? Well, Sun has worked with other major OS vendors to make Java work properly so I don't know what the issues are behind Linux and *BSD not doing the same. wonderful reasons why Sun only supports certain OS's snipped That's all lovely, I still don't have a recent version of Java on my machine. I don't expect this situation to really be fixed anytime soon. *As a Solaris gal, this might all seem perfectly sensible. Myself, I'm *doubly screwed running Linux (which is not Windows or Solaris) on a *PowerPC (which is not Intel or Sparc). Sun might have perfectly valid *business reasons for concentrating on Solaris and Windows, but in the *end it still means I'm screwed. There are platforms which are also difficult to install Perl though it has gotten better over the years. I remember when I first installed or rather tried to install perl5 on Solaris when it first came outit was a PITA to say the least. P5P was conceived for just this task to make Perl install and run on a wide variety of platforms since, at the time, the number was few. Nothing installs perfectly on every OS and runs perfectly as well. It such a thing exists, I'll buy stock in the company right now :) Yeah, but Perl does an amazing job of trying. About the only thing I can think that's been so well ported is maybe something like gcc, vi or make. It runs on the Amiga! It supports EBCDIC for god's sake! I think the key difference here is that p5p is committed to trying like hell to get Perl working on pretty much everything, while Sun is only doing the ones it thinks are important and leaving the rest to independent porting operations it gives its blessing, and source code, to (like Blackdown). Yes, you still need a White Knight for your OS, yes things take time to firm up, yes its not perfect, but we're trying and succeeding. Sun isn't even trying. Java's been around long enough and sure as hell has had enough resources poured into it. If they really wanted Java to be truely cross-platform, they'd have done it by now. They don't care. That's fine as a business, but like I said, whatever the reason the end-users are still fucked. *Given that Sun charges $200 per incident for a two day response time *($1600 for four hour reponse) this is not something most people can *afford. Expensive, commercial support, while it is nice for those who *can afford it, is not a general solution. When you have web sties that loose millions of dollars per hour of down time it's not an option not to have support. Sun is a business, Perl is not. People don't answer pagers at 4am for free :) You misread. Commercial support is entirely necessary and justifiable in your case and many others. It must exist. However, you can't hold it up as a good general solution. Advocating, When I had a problem, Sun fixed it right away mumblebecause I pay them lots of money/mumble is cheating. Take any technology, apply lots of money, and it will work wonderfully. It would be neat to have 'make solaris-dist' or something like that for most of the major platforms. ... NAPC sounds cool but I don't know enough about it to say whether or not it will be Enterprise level. Given that Enterprise has pretty much no meaning anymore, I'd say yes. :) Seriously though, NAPC (if it works) is exactly the sort of packaging tool you're looking for. Since NAPC will produce, essentially, a binary package, you can just translate that into a solaris package. Anyhow, I think we've beaten this topic to death. Point is, Perl's installation isn't any worse than Java and possibly better. -- Michael G. Schwern [EMAIL PROTECTED]http://www.pobox.com/~schwern/ Perl6 Quality Assurance [EMAIL PROTECTED] Kwalitee Is Job One Schwern Unit: a positive but insignificant quantity