Re: Perl, the new generation
I read *all* of Camel 1, it was a slim volume and gave (correctly) the impression that Perl is an easy to get into language that is useful for loads of things. You can justifiably still say those things about Perl -- but it's not the impression you get; and the impression is a lot of what counts. Mike - Original Message - From: Nick Stankus [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Friday, May 18, 2001 5:49 PM Subject: Re: Perl, the new generation Someone looking at that is going to think they have to know all that to be effective. Who reads the book. I just use it as reference. I am not the best Perl guru in the world, but I can program everything I need perl to do. If I ever need help...it is back to the Perl Camel Book. 2nd edition. or even Learning Perl works good ... if the problem is forgetfulness. nixter out. www.angryshirts.com -- piss people off.
Re: Perl, the new generation
Adam Turoff [EMAIL PROTECTED] writes: It's also amazing how long some people can go without seeing a statement modifier or non-default delimiters like s{}{};. In the micro view, that's OK. In the macro view, it leads to Perl Mongers meetings that feel more like AA: Which reminds me, must write up that proposal for YAPC::Europe... -- Piers Cawley www.iterative-software.com
Re: Perl, the new generation
Lightning flashed, thunder crashed and Russ Allbery [EMAIL PROTECTED] whispere d: | All Perl programmers, including lone ones, really should be using CPAN as | much as they can, which means that the parts of the language needed to use | CPAN modules are part of the understanding you need. This comment in and of itself sets a very high bar on perl's usability. In essense, you are saying that to use perl you must know (of) the 2500+ modules in CPAN. You are also saying that OOP is now required, because many/most CPAN modules use OOP. -spp
Re: Perl, the new generation
What is Camel4 going to look like for perl 6? What is going to be required knowledge for perl6. Let's just start by looking at Apoc2. To use perl, you'll have to know Unicode, you'll have to know OO, you'll have to understand references. Those are three very technical concepts that make Ummm, I must have missed the have to know Unicode, have to to know OO, have to know references part in the Apoc2. Could you show it to me? using perl to quickly throw things together much more difficult. And that's just Apoc2 -spp -- $jhi++; # http://www.iki.fi/jhi/ # There is this special biologist word we use for 'stable'. # It is 'dead'. -- Jack Cohen
Re: Perl, the new generation
Stephen P. Potter writes: For example, take a look at Camel1. It was a small book; you could carry it around without building up huge biceps. You could reasonable read it in a couple of days and get started with perl. I tried to get us to maintain that in Camel2, but it grew to almost 700 pages. Camel3 is 1100 pages, about a 3 fold increase from Camel1. I can weightlift with it now. Someone looking at that is going to think they have to know all that to be effective. Measuring the complexity of Perl by looking at the size of the Camel is bogus. The Camel is a reference book. Most of the bulk comes from describing things *better*. The new stuff in Perl between Camels 2 and 3 is minute. The book difference between Camels 2 and 3 is almost entirely made up of better explanations (e.g., modules and objects), more information (e.g., regexps described in detail for the first time, the additional function information in the perlfunc part), and so on. Nat
Re: Perl, the new generation
Lightning flashed, thunder crashed and Trond Michelsen [EMAIL PROTECTED] whis pered: | You don't need to know any of the modules in CPAN to use perl, but once | you learn how to use search.cpan.org, your productivity will most | probably increase dramatically. Just like knowing how to use the | documentation will make you more productive. The problem with this philosophy, is that you are talking about a fullblown software development situation. You are probably correct if I want to write a huge application. If I just want to write a quick script to (for example) do disk usage reporting. This was my first perl script, almost 10 years ago now. I was able to go from 0 (picking up Camel1) to having this script completed and working in 2 days. I just randomly chose 4 modules from the Security category. Two were dead links (User::pwent, User::grent), one (MD5) was a wrapper that sent me elsewhere (Digest::). It has both a functional interface and an OO interface. The last has only an OO interface (Authen::ACE). While I was there I (hypothetically) decided I wanted to write a web page. I searched for Web. 112 packages in 24 distributions. To look through all that is going to take a lot more time than I want to spend on writing a web page. I'll just do it by hand. I was also thinking of doing some CGI. 369 modules in 81 distributions. 75% (approximately) have no synopsis listed, so I have no idea except by name what they do. -spp
Re: Perl, the new generation
Someone looking at that is going to think they have to know all that to be effective. Who reads the book. I just use it as reference. I am not the best Perl guru in the world, but I can program everything I need perl to do. If I ever need help...it is back to the Perl Camel Book. 2nd edition. or even Learning Perl works good ... if the problem is forgetfulness. nixter out. www.angryshirts.com -- piss people off.
Re: Perl, the new generation
Stephen P. Potter writes: | You don't need to know any of the modules in CPAN to use perl, but once | you learn how to use search.cpan.org, your productivity will most | probably increase dramatically. Just like knowing how to use the | documentation will make you more productive. The problem with this philosophy ... This is off-topic for perl6. Nat
Re: Perl, the new generation
Lightning flashed, thunder crashed and Jarkko Hietaniemi [EMAIL PROTECTED] whispered : | Ummm, I must have missed the have to know Unicode, have to to know OO, | have to know references part in the Apoc2. Could you show it to me? Atoms- Unicode. If everything is Unicode, you're going to have to grok Unicode (at least tangentally) to be able to use perl. RFC 161- Everything becomes an object. Filehandles are more object oriented in Perl6, and the special variables So, *probably* if you are going to use filehandles, you'll have to grok OO. $#foo is gone. If you want the final subscript of an array, and [-1] isn't good enough, use @foo.end instead. There's *lots* of mention of OO in relatively common, normal things. RFC 009- That is, all variables may be thought of as references, not just scalars. The whole concept of $calar, @rray, and %ash are changing to become references. If you don't understand references, you won't be able to use variables. That pretty much seems to say to mee you must know OO and references. Unicode may be something more easily hidden under the rug. -spp
Re: Perl, the new generation
Lightning flashed, thunder crashed and Nathan Torkington [EMAIL PROTECTED] whi spered: | This is off-topic for perl6. Objection, your honor! This is a logical extention of part of the discussion. If we're discussing what is wrong with perl5 to make perl6 better differentiating between philosophies is quite on target. -spp
Re: Perl, the new generation
On Fri, May 18, 2001 at 11:24:45AM -0400, Stephen P. Potter wrote: You are also saying that OOP is now required, because many/most CPAN modules use OOP. This is a piece of FUD along the lines of inline POD slows code down that keeps people fearful of CPAN and I'd really rather see die. To *use* OO code is a much different (and smaller) beast than *designing* OO code. For most OO CPAN modules, about all you need to know about OO is the syntax of calling a method. This should be obvious just from reading the module docs (assuming its well-documented). In all other respects it works just like its functional equivelent. There are exceptions (such as LWP) but most have simple wrappers (LWP::Simple) and who's to say their function-oriented equivalents would be any less complex? Sean Burke wrote up an excellent article about OO for module users which I thought was on perl.com but I can't find at the moment. Maybe it was in TPJ. -- Michael G. Schwern [EMAIL PROTECTED]http://www.pobox.com/~schwern/ Perl6 Quality Assurance [EMAIL PROTECTED] Kwalitee Is Job One Follow me to certain death! http://www.unamerican.com/
Re: Perl, the new generation
On Fri, May 18, 2001 at 12:22:56PM -0400, Stephen P. Potter wrote: For example, take a look at Camel1. It was a small book; you could carry it around without building up huge biceps. You could reasonable read it in a couple of days and get started with perl. I tried to get us to maintain that in Camel2, but it grew to almost 700 pages. Camel3 is 1100 pages, about a 3 fold increase from Camel1. I can weightlift with it now. Someone looking at that is going to think they have to know all that to be effective. Programming Perl is a reference manual. It is designed to cover the whole of the language in detail. It will be large. Learning Perl is the tutorial. It is designed to cover just the basics. It will be small. Trying to learn Perl from the Camel is like trying to learn English from the OED. Trying to make Perl easier to learn by cutting features is about as sensible as making English easier to learn by tearing pages out of the OED. The size of the language has little to do with its difficulty, its more about the minimum subset that must be learned to be useful (the *actual* subset, not the perceived). In fact, the Llama 2 (written for 5.004) doesn't cover much of perl5 at all. I don't think it ever mentions OO or references except in passing. Llama 3 should be much the same. What you are worried about is not a language issue, it is a perception and DOCUMENTATION issue. Please stop throwing your wooden shoes in the cogs of progress and start helping. Go to [EMAIL PROTECTED], start up a perlsmall man page. -- Michael G. Schwern [EMAIL PROTECTED]http://www.pobox.com/~schwern/ Perl6 Quality Assurance [EMAIL PROTECTED] Kwalitee Is Job One How can I stoop so low? Years of practise, that's how. It's been hard going but now I can stoop lower than a pygmy limbo dancer. -- BOFH
Re: Perl, the new generation
On Fri, May 18, 2001 at 07:16:36PM +0100, Michael G Schwern wrote: Sean Burke wrote up an excellent article about OO for module users which I thought was on perl.com but I can't find at the moment. Maybe it was in TPJ. http://search.cpan.org/doc/SBURKE/HTML-Tree-3.11/lib/HTML/Tree/AboutObjects.pod -- // Trond Michelsen \X/ [EMAIL PROTECTED]
Re: Perl, the new generation
On Fri, May 18, 2001 at 12:55:55PM -0400, Stephen P. Potter wrote: Atoms- Unicode. If everything is Unicode, you're going to have to grok Unicode (at least tangentally) to be able to use perl. Bah. Rubbish, no more than you need to grok Unicode to use Perl 5.6. Do you know what data of yours 5.6 is storing in Unicode? No. Do you care? No. Do you need to? No. All filenames in Windows 2000 are, I'm told, in Unicode now; I don't *think* that means that anyone who wants to use Windows 2000 has to grok Unicode. -- [It is] best to confuse only one issue at a time. -- KR
Re: Perl, the new generation
On Fri, May 18, 2001 at 08:08:40PM +0100, Simon Cozens wrote: On Fri, May 18, 2001 at 12:55:55PM -0400, Stephen P. Potter wrote: Atoms- Unicode. If everything is Unicode, you're going to have to grok Unicode (at least tangentally) to be able to use perl. Bah. Rubbish, no more than you need to grok Unicode to use Perl 5.6. Do you know what data of yours 5.6 is storing in Unicode? No. Do you care? No. Do you need to? No. One of the big selling points about Java is that it's always use Unicode natively from day 1, yet I've never seen a Unicode Primer for programmers starting out with Java book/site/article/paper/certification. Unicode is just *there*. Much like oxygen and nitrogen. The tangential deviation necessary to grok unicode to use Perl is perhaps .01 degrees away from the previous learning curve. Using Perl to grok Unicode is a little different. :-) Z.
Re: Perl, the new generation
Stephen P. Potter writes: Objection, your honor! This is a logical extention of part of the discussion. If we're discussing what is wrong with perl5 to make perl6 better differentiating between philosophies is quite on target. The corner of the discussion about search.cpan.org and broken modules didn't seem particularly philosophical. Nat
Re: Perl, the new generation
Stephen P. Potter writes: Atoms- Unicode. If everything is Unicode, you're going to have to grok Unicode (at least tangentally) to be able to use perl. Others have well dealt to this. RFC 161- Everything becomes an object. Filehandles are more object oriented in Perl6, and the special variables So, *probably* if you are going to use filehandles, you'll have to grok OO. $#foo is gone. If you want the final subscript of an array, and [-1] isn't good enough, use @foo.end instead. There's *lots* of mention of OO in relatively common, normal things. You're partially correct, in that more things will be becoming OO. I'm picking that the old special variables and select() were harder to grok than the new filehandle OO doodads will be. RFC 009- That is, all variables may be thought of as references, not just scalars. The whole concept of $calar, @rray, and %ash are changing to become references. If you don't understand references, you won't be able to use variables. Once again, partially correct but with some FUD. RFC 9 wasn't accepted in full. %foo, @foo, and $foo are not the same thing. What was taken was $ref = %foo; makes $ref hold a hash reference. If you're a beginner and don't know about references, don't do that. It makes reference stuff more convenient for those who know how to do it, and also simplifies some of the wackier prototyping. This is hardly if you don't understand references, you won't be able to use variables. Nat
Re: Perl, the new generation
On Wed, May 16, 2001 at 11:58:07AM -0400, Adam Turoff wrote: It's not so much that Perl shouldn't have data structures or modules. I think what Stephen is saying (and he's not the only one) is that the bare minimum amount of Perl you *must* know to be productive is increasing. Either that, or we're giving the impression that it's increasing. This may have gotten lost in the noise, so I'll mention it again. Since all the features of perl4 are still in perl5 (mod a few minor differences) it should still be possible to teach perl5 as you did perl4, as a small utility language/shell scripting replacement. Simply ignore anything that might get in the way of someone just wanting to read a log file. We could provide a seperate man page (perlsmall?) which describes this mini-language-within-a-language. It would skip things like OO (or only as much as you need to use the occasional CPAN module), Unicode, odd syntax details, etc... and focus on things like basic regexes, string and file handling, basic data structures, map, grep, etc... In fact, you could start with the perl4 man page (or Camel/Llama 1) as the basis. Seems a lot more productive than just pining about the olden days when men were men, Perl was small and 640K was enough for anyone. -- Michael G. Schwern [EMAIL PROTECTED]http://www.pobox.com/~schwern/ Perl6 Quality Assurance [EMAIL PROTECTED] Kwalitee Is Job One ...and I pull out the Magnum from under the desk where I keep it in case someone laughs at a joke that's so dry it's got a built in water-fountain, and blow the lot of them away as a community Service. -- BOFH
Re: Perl, the new generation
Hmmm...ok, on thinking about it, I generally agree with you. There is only one point that I would debate (and, as you'll see, there's a solution for that one, too): On Wed, 16 May 2001, Nathan Torkington wrote: Dave Storrs writes: 1) One of the great strengths of Perl is that its learning curve is very shallow but very long. Adding more stuff to the language makes the curve steeper, because you need to hold more in your head as you learn it. I see those as orthogonal. I can add more to the high end of a language that beginners don't need to know. While it may be true that beginners don't need to use a particular feature--or even know about it--how will they know that until they have studied it? ACTION = insert($tongue, $cheek) Imagine the following conversation: JAPH: Here's a list of all the features in Perl. It may look overwhelming, but don't worry...you don't need to know all of them until later. Beginner: GAACKK!!! /ACTION Actually, something like what Randal was recently talking about, with the llama (i.e., introduction, small subset, whathaveyou) probably addresses this concern. We just have to make sure to point everyone at that document as soon as possible upon their entry into Perl. Dave
Re: Perl, the new generation
LOL! No bias there then Nat :-) Mike - Original Message - From: Nathan Torkington [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, May 15, 2001 10:41 PM Subject: Re: Perl, the new generation Stephen P. Potter writes: It seems to me that recently (the last two years or so) and [and other stuff] The perl6 runtime will be separate from the language parser, so you could write a perl4 parser to run on the perl6 runtime if you wanted to be so perverse. Nat
Re: Perl, the new generation
On Thu 17 May, Michael G Schwern wrote: On Wed, May 16, 2001 at 11:58:07AM -0400, Adam Turoff wrote: It's not so much that Perl shouldn't have data structures or modules. I think what Stephen is saying (and he's not the only one) is that the bare minimum amount of Perl you *must* know to be productive is increasing. Either that, or we're giving the impression that it's increasing. We could provide a seperate man page (perlsmall?) which describes this mini-language-within-a-language. It would skip things like OO (or only as much as you need to use the occasional CPAN module), Unicode, odd syntax details, etc... and focus on things like basic regexes, string and file handling, basic data structures, map, grep, etc... Many years ago I had dealings with a language which was effectivly documented in a two diminsional manner. It had a very brief introduction, then you could read the first section of each chapter to get a language over view - enough for simple use, then if you needed the depth you read the other sections. The book even had a two dimensional contents matrix at the front. This worked (for me). Richard -- [EMAIL PROTECTED]
Re: Perl, the new generation
On Tue, May 15, 2001 at 03:41:15PM -0600, Nathan Torkington wrote: Stephen P. Potter writes: It seems to me that recently (the last two years or so) and especially with 6, perl is no longer the SAs friend. It is no longer a fun litle language that can be easily used to hack out solutions to problems. It is now (becoming) a full featured language, quite at the expense of its heritage. And yet there are a zillion programs from perl4 and earlier that still work in perl5. In what way can you not use Perl to solve sysadmin problems or hack out fun solutions to problems? I do those two things all the time. I don't think backwards compatibility is the point here. I picked up Camel 1 recently, and it was quite amazing how different Perl4 *felt*. It's like Perl was being pitched as a good language for writing standalone programs or utilities of standalone programs (the type a sysadmin would use). It didn't feel like it was being offered as the kind of language to write things like Class::Contract, Inline::C, AxKit, SOAP::Lite or the all-singing-all-dancing CGI.pm. Where are we now? Perl5 is a bigger language and Perl6 is proposed to be bigger still. There are people who complain about Perl5 because they can't keep it all in their heads, unlike C, sh and Python (and to some extent, Perl4). When we moved from 4 to 5, so people thought we should continue developing 4 without all the useless new stuff, like OO and threads and etc. I wonder more and more if they weren't right. I wonder if as 6 develops if we shouldn't split off the old 4 syntax and have two languages. If you want to do it, do it. I vomit at the thought of a language without data structures or modules, though, and I wouldn't be surprised if others did too. It's not so much that Perl shouldn't have data structures or modules. I think what Stephen is saying (and he's not the only one) is that the bare minimum amount of Perl you *must* know to be productive is increasing. Either that, or we're giving the impression that it's increasing. Many people don't want to get bogged down in how the details of Unicode, upperclass level CS topics or Perl's unique syntactical peculiarities to parse a damn log file (or find and use a CPAN module that does it). Z.
RE: Perl, the new generation
On Wed, 16 May 2001, David Grove wrote: For me, it's the bare minimum amount of Perl you must *use* to be productive that I see increasing in our plans and discussions. I'm afraid of Perl turning into a verbose monstrosity to please verbosity addicts of languages whose only point of advocacy is Perl FUD. Once quick and dirty dies, Perl dies. Several thoughts for you, David. All of these should be taken from the perspective of someone who cut his teeth on 5.x and has never had to deal with the (joys|differences|horrors) of 4.x. 1) I agree that Perl is a big language and it's hard to hold it in your head. I frequently find that some bit of it that I haven't used in a while has fallen out and I need to go read up on it again. 2) Respectfully, I don't think that we can accurately say that the minimum amount of Perl needed in order to be productive is increasing; we haven't finished defining P6 yet, so how can we know this? 3) You have every right to be afraid of anything you want to be afraid of, and to express your concerns about it. However, the way that you chose to do that (Once quick and dirty dies, Perl dies.) implies that the only thing that Perl is good for is q-n-d, and this is simply not the case. I have written enterprise-quality code, for large systems, in Perl, and I will absolutely defend Perl's ability on that playing field. 4) While your concern is well taken, I think you are doing yourself a disservice by using such inflammatory language...it makes me (and probably others) focus more on your tone than on your point. Dave Storrs
Re: Perl, the new generation
On Wed, May 16, 2001 at 11:14:57AM -0700, Dave Storrs wrote: afraid of, and to express your concerns about it. However, the way that you chose to do that (Once quick and dirty dies, Perl dies.) implies that the only thing that Perl is good for is q-n-d A veritable lesson in logic! Here's an equivalent statement. Once all the oxygen suddenly disappears from the atmosphere, humanity is wiped out. That naturally suggests that the only thing humanity is good for is is respiring oxygen, right? And it's an almost *exactly* equivalent statement, because it's almost as likely that Perl will stop being good for quick 'n' dirty stuff as all the oxygen dropping out of the atmosphere. -- Twofish Pokemon seems an evil concept. Kid hunts animals, and takes them from the wild into captivity, where he trains them to fight, and then fights them to the death against other people's pokemon. Doesn't this remind you of say, cock fighting?
Re: Perl, the new generation
On Wednesday 16 May 2001 15:32, Nathan Torkington wrote: Bryan C. Warnock writes: I think the biggest fear isn't that Perl is going to grow out of its niche, but that it's going to outgrow it. It's great that Perl has been able to expand to be so many things to so many people, but not at the expense of forgetting its roots - of the whole Right Tool / Right Job that it came from. In that case, how exactly has it forgotten its roots? I mean, in what way is it not as useful as it was before? Nat Sorry. I didn't mean to imply that it had, only that it seems the largest fears center around that it will. Certainly, we are doing our best to keep Perl Perl. But in the process of overruning enemy camps, are we leaving our own camp unguarded? One of the nice things about early Perl 5 (I'm sorry - I was crawling through mud for most of Perl 4) is that Perl was an additive language. You had simple concepts to accomplish simple tasks. As the tasks got more complex, you could add more complex concepts onto your simple knowledge base to accomplish them. Of course, when writing RFCs, no one (except for Keep Perl Perl) really addressed what is right with Perl, or Why Things Were Good. People addressed ways that Perl *could* be improved, with the hopes that Larry would be able to differentiate between 'what would make a better language', and 'what would make Perl better'. Those are two orthogonal concepts, when you think about it. So, in reading the RFCs, and in discussions centered around make a better language, Perl - at least, the old, simple Perl - has seemingly become a subtractive language. (Everything's an object, use warnings and strict by default, etc, etc) Perl would have a much higher barrier to entry for what used to be a simple task. What previously required the bare minimum of Perl knowledge to do, would now require more complex conceptual issues, if only to determine what extraneous features can be removed or hidden, and how to do that. -- Bryan C. Warnock [EMAIL PROTECTED]
Re: Perl, the new generation
Dan == Dan Sugalski [EMAIL PROTECTED] writes: Dan People think they *must* know all the core bits of a language, and Dan they think that consists of all the stuff we ship with perl. (And, Dan let's face it, we ship a *lot* of stuff with perl) It's like you're Dan not allowed to know only a part of a language anymore--that's somehow Dan ungeeky or something. One of the things that I think made llama1 so successful and our ongoing llama course popular is that I deliberately chose a subset of Perl to teach that was reasonably self-consistent and yet covered 80% of people wanted to do with Perl in under 100 lines of code. And that's not an easy task. I thought long and hard about where to put array slices and alternate quoting operators. I had to think about how much new problem space a given feature opens for the number of paragraphs it would take to describe the feature. That's why you don't see heredocs in llama1. It's easy to describe for those familiar with the Unix shell, but if you're talking about a standalone tutorial, it takes a good page or two to really talk about the nuances. And the same problem space was covered fine by just letting a double-quoted string break over many lines. :) And that's why you still won't see heredocs in llama3. The point of the llama is to cover a subset that lets you implement most of what you need for most of the programs out there. In fact, Tom Phoenix and I killed formats. Yeah, people use them, but the 10 pages for that were sorely needed to cover other things in more detail or more breadth. So, when Perl6 stabilizes a bit more, I'll be starting the same process for Perl6 what's the tiny language buried within this large language? I bet I can still give you the first 40 hours that everyone needs starting with Perl6 in a 250-page book, and it'll still cover 80% of what everyone needs for 80% of the programs. I bet about 75% will be the same semantics as llama3, with some syntax changes. -- Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 [EMAIL PROTECTED] URL:http://www.stonehenge.com/merlyn/ Perl/Unix/security consulting, Technical writing, Comedy, etc. etc. See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl training!
Re: Perl, the new generation
On Wed, 16 May 2001, Nathan Torkington wrote: Dave Storrs writes: SARCASM=EXTREME Everyone, please try to stop the downhill descent of the conversation. This is not just Dave, but others in the thread too. For the record, the original post in this sequence came from David Grove, not from me (David Storrs). My response to David was an attempt at *preventing* a downhill descent...which is why Simon's comment, which came off feeling abrasive to me, bothered me. You're right; I should have refrained from sarcasm and simply asked Simon to please not treat my concerns so dismissively. It sounds like the concern is that each new version of Perl adds features, which programmers use. To be able to maintain or extend code, you need to know those features. Thus, the core knowledge for survival in Perl, is ever-growing. This is what I understood to be David Grove's point (David, please correct me if I have misunderstood). I don't know if I agree with this (I also may not have the background to answer it, since I didn't come on until 5.x), but I do feel, as I said before, that the language is sufficently large that it is hard to hold in one's head and that making it significantly larger would be a cause for concern. Other people may disagree with me on this; it's only my opinion. In some ways I agree with this. In particular, the growing number of modules with an OO interface means that knowing how to use objects is more and more important. This is true, but it could be taken as a counterargument...if there is a growing number of OO modules, that is because a growing number of Perl programmers are accustomed to, and make use of, OO techniques. [single programmer doesn't need advanced features, teams are not used for solving small problems so it is reasonable that they need advanced stuff] So I guess I don't see it as that big a problem. Am I missing something? Well...I'm not sure my concerns are well enough defined to be convincing, but I'll try to lay them out: 1) One of the great strengths of Perl is that its learning curve is very shallow but very long. Adding more stuff to the language makes the curve steeper, because you need to hold more in your head as you learn it. 2) If the language is so big that you can't hold all of its features in your head, then those extra features might as well not exist. Now, after all of the above discussion, I should just say that I'm not convinced that Perl is too big (I think it's _big_, which is different from _too_ big), or that anything that we are adding is going to _make_ it too big. I'm simply trying to point out one side of the argument. Dave Storrs
Re: Perl, the new generation
Lightning flashed, thunder crashed and Larry Wall [EMAIL PROTECTED] whispered: | Peter Scott writes: | : So, I wonder aloud, do we want to signify that degree of change with a more | : dramatic change in the name? | | I'm inclined to think that people will be more likely to migrate if | they subconsciously think we're taking continuity into consideration. | Which we are, albeit not at a syntactic compatibility level. It seems to me that recently (the last two years or so) and especially with 6, perl is no longer the SAs friend. It is no longer a fun litle language that can be easily used to hack out solutions to problems. It is now (becoming) a full featured language, quite at the expense of its heritage. When we moved from 4 to 5, so people thought we should continue developing 4 without all the useless new stuff, like OO and threads and etc. I wonder more and more if they weren't right. I wonder if as 6 develops if we shouldn't split off the old 4 syntax and have two languages. -spp
perlsmall (was Re: Perl, the new generation)
On Tue, May 15, 2001 at 03:01:47PM -0400, Stephen P. Potter wrote: It seems to me that recently (the last two years or so) and especially with 6, perl is no longer the SAs friend. It is no longer a fun litle language that can be easily used to hack out solutions to problems. See, I have a basic problem with this. Whatever you were doing with perl4 ten years ago, you can still do with perl 5.6.1 (mod a few minor differences) by simply ignoring all the new features. Perl can be taught/learned as a small, fun language (where 'fun' is a highly relative term). In fact, there's an O'Reilly book out just for that purpose, Perl For System Administration. While there are arguments about unnecessary language bloat and what should be in the core and what should be relegated to modules (see also, Second System Effect) if you take this too far you wind up with a Luddite philosophy to language design. I don't need feature X, so neither does anyone else! There's little need for a language fork. The simple things will remain simple (and hopefully simpler) and the hard things will remain possible (and hopefully a bit more possible). Perhaps instead of proposing a fork, you could write up a new man page, perlsmall or something, describing just those useful features of Perl as a small utility language (enhanced shell scripting) and ignoring the useless stuff like OO and threads. -- Michael G. Schwern [EMAIL PROTECTED]http://www.pobox.com/~schwern/ Perl6 Quality Assurance [EMAIL PROTECTED] Kwalitee Is Job One Let's enjoy the traditional custom in Peru of getting leprosy.
Re: Perl, the new generation
David Grove [EMAIL PROTECTED] writes: Perl 5 is far from stagnant--please don't bend the truth to fit your points. My impression is that there's quite a bit more constructive activity on p5p than there was a year ago. I've stopped paying attention to P5P except for keeping an eye on the possibility of a new surprise upgrade from Microsoft. However, the attitude of the P5P is irrlevant to the user base. : Unless Perl 6 is capable of parsing and running that 99.9% (or higher) of : Perl 5 scripts originally foretold, I foresee a far worse outcome for Perl 6 : than has happened for an almost universally rejected 5.6 and 5.6.1. There you go again, as Uncle Ronnie used to say. Excessive hyperbole will cost you sympathetic readership. Shall I list them again? Dude, it's been 13 months since 5.6 was released, and two commercial entities have so far accepted it: ActiveState and SuSE. Speaking with SuSE around October (7.0), the rep's answer getting back to me was simply we don't consider it to be stable enough yet to include it in our distribution. Well, it's there in Mandrake 8, and was available as an update long before Mandrake 8 got released. Still only 5.6.0 though. Dunno about the rest, but it's 50% more than your claim... And remind me how long ago it was that most of the systems you're talking about actually started to include Perl as anything other than a 'Danger Will Robinson, unsupported contrib code' type package? -- Piers Cawley www.iterative-software.com
Re: Perl, the new generation
Hey, we could call it Perl 9 from Outer Space. No wait... Larry
Re: Perl, the new generation
At 05:36 PM 5/10/01 +0100, Michael G Schwern wrote: Version numbers are, at best, an indication of the magnitude change. At worst they are a cheap marketing ploy. I've always liked that Perl's version numbers are relatively free of marketing hoopla (the jump from perl3 to perl4 notwithstanding). The move from 5.005_03 to 5.6.0 style was jarring enough (and fairly well justified). Its been so long since we've had an integer increment that it should be fairly shocking. Eh, I fully understand that version number magnitudes are simply to attract attention, and that The Faithful don't need the glitz. Since AFAICT the glitz doesn't hurt, though, it doesn't do any harm to give marketing all the help it can get; and let's face it, marketing hasn't been Perl's greatest strength. I was one of the people calling for 5.006 - 5.6, since the changes, to me, were greater than what was implied by an increment in the fourth significant digit. And it worked, too; I finally saw a couple of articles in trade (non-geek) rags about the upgrade. More or less the only articles about Perl I've seen there for 5 years. (I'm talking about rags like Information Week, Internet Week, Computerworld, that sort of thing.) I'm just applying the same principle here, comparing to the Perl 4 - Perl 5 change. Like I said, I figure it's a long shot; I just thought I'd run it up the flagpole. -- Peter Scott Pacific Systems Design Technologies http://www.perldebugged.com
RE: Perl, the new generation
I've been wondering for quite some time whether we were creating a Perl for the purpose of cleaning up the ridiculously rigged Perl 5 internals, or creating a Perl for the simple enjoyment of creating a new programming language. Certainly, recent discussions would point to the latter; as we move farther and farther away from Perl 5 syntax, we move dangerously close to completely closing Perl as a viable tool for the gazillions of users who have the misfortune of legacy. This legacy isn't just a website or a utility here and there anymore, but often an entire suite of software, or tools integrated into operating systems, some or much of which the user may not even be aware of. Translating is not an option for these people. A slow transition may be a catchphrase nowadays, but with Perl 5 stagnant, 5.6 accepted on only two systems that I'm aware of (SuSE and Win32/AS; rejected everywhere else), and PHP/Python/.NET ready to swollow up anyone who would believe anything, I'm concerned that this transition may not exist. So, I'll go you one farther. What about creating a cleaned up perl, and letting those who want to play with a new language entirely do so in the form of a true fork. Certainly, Perl 6 is coming to resemble Perl 5 little more than PHP and resembles Perl 5 and Perl 5 resembles C. We haven't even started writing the actual tool(s): we haven't even completed planning without coming up with a tool that only resembles Perl due to a use of $@%, as an offspring rather than a serious hot bath. If we keep this up, Larry's 95% mark will end up going to 90%, 85%, and then who knows. I DO NOT DISLIKE the changes that I'm seeing. However, their coolness ends when it comes time to trace through my entire operating system(s) and change every perl file that exists here; and the thought of a mass exodous to Python/PHP because we've made Perl 5 obsolete and scared off the rest of our community, especially corporate members, is completely unappealing. Corporate users do not think in terms of neat and novel, they think in terms of how much work it's going to be to keep up with the complete overhaul of a language versus moving to a language with a stable syntax once and not having to deal with it again. We will not soon rise above that kind of bad opinion. FUD? Perhaps. Reality? Definitely. Python books are already full of FUD, and I've had to stop reading .NET books because just holding the books in my hand makes my blood pressure rise 90 points. Imagine what will happen when that FUD turns serious and actually costs Perl users a great deal of money? Unless Perl 6 is capable of parsing and running that 99.9% (or higher) of Perl 5 scripts originally foretold, I foresee a far worse outcome for Perl 6 than has happened for an almost universally rejected 5.6 and 5.6.1. Fun is fun. But work costs money, guys. And if you cost people money with a free tool, repercussions could be bad not just for Perl but for free languages, among which Perl has heretofore been the leader of the pack. Actually, Peter, I was getting very, very close to writing this anyway. David T. Grove Blue Square Group [EMAIL PROTECTED] [EMAIL PROTECTED] -Original Message- From: Peter Scott [mailto:[EMAIL PROTECTED]] Sent: Thursday, May 10, 2001 12:20 PM To: [EMAIL PROTECTED] Subject: Perl, the new generation This is a long shot, but here goes. I was thinking about Perl 6 this morning while jogging (blithely ignoring the forest scenery). It occurred to me that what appears to be emerging as the new language embodies bigger changes than I ever anticipated - which is great, software should improve with time. And so I found myself wondering whether the title does it justice. Perl 6 is looking to me almost like an entirely new language. The change from Perl 5 to Perl 6 is much, much larger than the change from Perl 4 to Perl 5 (virtually all Perl 4 code ran unmodified under Perl 5). So, I wonder aloud, do we want to signify that degree of change with a more dramatic change in the name? Still Perl, but maybe Perl 7, Perl 10, Perl 2001, Perl NG, Perl* - heck, I don't know, I'm just trying to get the creative juices flowing. I do believe that the tremendous effort that is going into Perl 6 deserves more attention than I think it will get with that title. At some point, the Perl 6 cognomen will have attracted enough inertia that we couldn't reasonably change it even if we wanted to. Maybe that time has already come. Maybe not. Can't hurt to raise the question. -- Peter Scott Pacific Systems Design Technologies http://www.perldebugged.com
RE: Perl, the new generation
Incompatible continuity. Sounds like Microsoft marketing. We're strongly considering keeping compatibility, and rejecting it wherever we can insert something that looks momentarily cool. Of course your Perl 5 programs will still work, as long as you convert them to Perl 6. We'll have a parser that will be able to do this. Of course, you will have to write it yourself. Perl 6 will still be perl, because the name won't change... the language is a different matter entirely. Doesn't wash... A non-MS-microweenie can only digest a limited number of oxymora at a time. David T. Grove Blue Square Group [EMAIL PROTECTED] [EMAIL PROTECTED] -Original Message- From: Larry Wall [mailto:[EMAIL PROTECTED]] Sent: Thursday, May 10, 2001 12:44 PM To: Peter Scott Cc: [EMAIL PROTECTED] Subject: Re: Perl, the new generation Peter Scott writes: : So, I wonder aloud, do we want to signify that degree of change with a more : dramatic change in the name? I'm inclined to think that people will be more likely to migrate if they subconsciously think we're taking continuity into consideration. Which we are, albeit not at a syntactic compatibility level. Larry
Re: Perl, the new generation
On Thu, May 10, 2001 at 12:56:36PM -0400, David Grove wrote: Of course your Perl 5 programs will still work, as long as you convert them to Perl 6. We'll have a parser that will be able to do this. Of course, you will have to write it yourself. I think there's a communications foul-up here. We're definately providing some sort of Perl 5 translator/adaptor system and we're definately writing it. In fact, it was hotly debated on perl6-language just recently exactly how to do it cleanly. Have I missed something? Have you missed something? -- Michael G. Schwern [EMAIL PROTECTED]http://www.pobox.com/~schwern/ Perl6 Quality Assurance [EMAIL PROTECTED] Kwalitee Is Job One mendel ScHWeRnsChweRNsChWErN SchweRN SCHWErNSChwERnsCHwERN sChWErn ScHWeRn schweRn sCHWErN schWeRnscHWeRN SchWeRN scHWErn SchwErn scHWErn ScHweRN sChwern scHWerNscHWeRn scHWerNScHwerN SChWeRN scHWeRn SchwERNschwERnSCHwern sCHWErN SCHWErN sChWeRn
Re: Perl, the new generation
Nathan Wiger writes: : Maybe the name Perl should be dropped altogether? No. The Schemers had to do a name change because the Lisp name had pretty much already been ruined by divergence. : (Granted, that's not what I'd prefer, but the changes are getting : rather massive and are starting to really permute the proposed : language) If you talk that way, people are going to start believing it. The typical Perl 6 program is not going to look very different from the typical Perl 5 program. The danger of us continually talking about the things we want to change is that people will forget to notice the tremendous amount of stuff that we aren't changing. Larry
Re: Perl, the new generation
On Thu, May 10, 2001 at 11:55:36AM -0700, Larry Wall wrote: If you talk that way, people are going to start believing it. The typical Perl 6 program is not going to look very different from the typical Perl 5 program. The danger of us continually talking about the things we want to change is that people will forget to notice the tremendous amount of stuff that we aren't changing. It might be useful to draw up a list of functions and features which we don't plan on changing? Maybe just run through each Perl 5 man page and highlight everything that will still be the same and post this somewhere? -- Michael G. Schwern [EMAIL PROTECTED]http://www.pobox.com/~schwern/ Perl6 Quality Assurance [EMAIL PROTECTED] Kwalitee Is Job One The desired effect is what you get when you improve your interplanetary funksmanship.
Re: Perl, the new generation
If you talk that way, people are going to start believing it. The typical Perl 6 program is not going to look very different from the typical Perl 5 program. The danger of us continually talking about the things we want to change is that people will forget to notice the tremendous amount of stuff that we aren't changing. Maybe, but for one I'm starting to wonder. TomC's rant rang true in my ears. How much can we change and still call it the same language? I'm not yet panicking, I'm just trying to hug some firm ground here. Larry -- $jhi++; # http://www.iki.fi/jhi/ # There is this special biologist word we use for 'stable'. # It is 'dead'. -- Jack Cohen
Re: Perl, the new generation
* Larry Wall [EMAIL PROTECTED] [05/10/2001 11:57]: Nathan Wiger writes: : Maybe the name Perl should be dropped altogether? No. The Schemers had to do a name change because the Lisp name had pretty much already been ruined by divergence. : (Granted, that's not what I'd prefer, but the changes are getting : rather massive and are starting to really permute the proposed : language) If you talk that way, people are going to start believing it. The typical Perl 6 program is not going to look very different from the typical Perl 5 program. The danger of us continually talking about the things we want to change is that people will forget to notice the tremendous amount of stuff that we aren't changing. Don't get me wrong - I'm not trying to be melodramatic. Far from it. However, one thing I worry we're losing sight of is *programmer* migration. We can write all the translators we want, but the person still has to learn Perl 6. As long as we're getting clear bang for the buck, then we're probably ok. But I continue to become increasingly worried that we're on a slippery slope of changes that really aren't needed. This may sound blasphemous, but I think we should try to change as few things as possible. By that I just mean let's determine what really needs to be overhauled. The $@% system? Yes. Apoc2 gets an A+ there. Bareword filehandles? Absolutely. Better semantics for passing @ around? Yup. But I think we just need to realize that every change we make is a change that thousands (millions?) of Perl programmers must now relearn. Since Perl is all about being programmer-centric, I think we just need to bear this in mind more closely when considering changes to such fundamental tenets as FILE and such. I would bet many JAPHs don't even know that you can say readline(FILE). -Nate
RE: Perl, the new generation
Perl 5 is far from stagnant--please don't bend the truth to fit your points. My impression is that there's quite a bit more constructive activity on p5p than there was a year ago. I've stopped paying attention to P5P except for keeping an eye on the possibility of a new surprise upgrade from Microsoft. However, the attitude of the P5P is irrlevant to the user base. : Unless Perl 6 is capable of parsing and running that 99.9% (or higher) of : Perl 5 scripts originally foretold, I foresee a far worse outcome for Perl 6 : than has happened for an almost universally rejected 5.6 and 5.6.1. There you go again, as Uncle Ronnie used to say. Excessive hyperbole will cost you sympathetic readership. Shall I list them again? Dude, it's been 13 months since 5.6 was released, and two commercial entities have so far accepted it: ActiveState and SuSE. Speaking with SuSE around October (7.0), the rep's answer getting back to me was simply we don't consider it to be stable enough yet to include it in our distribution. If Perl 5.6 hasn't caught on after a year, God help us trying to pass Perl 6 off as still Perl sometime this decade. Nobody ever foretold 99.9%, as far as I recall. I surely didn't. I was quoting you from about 5 messages ago... larryquote If 99.99% of scripts translate, that's good enough for me. :-) Actually, if 95% of Perl 5 scripts translate, I'll be overjoyed. /larryquote That's where it came from. The issue on the table is the magnitude of the diversion from Perl 5 to Perl 6, and my issue is its effect on the user base, especially commercial users and their legacy. I have to be concerned for these people: they're my customers and my peers. Perl 4 - Perl 5 happened at a time when perl wasn't considered THE scripting language. Entire commercial systems weren't widely written in it. Python, PHP, ASP, VBS, Ruby, and .NET weren't hot on it's tail spreading FUD to mezmerize users with transparent and ephemeral fancies. The effect on Win32 alone could be disastrous, so consider what would happen on systems where parts of the system itself were Perl 5 (some Linux distros lean heavily on Perl). p
Re: Perl, the new generation
- Original Message - From: David Grove [EMAIL PROTECTED] To: Peter Scott [EMAIL PROTECTED]; [EMAIL PROTECTED] Sent: Thursday, May 10, 2001 5:47 PM Subject: RE: Perl, the new generation . . . Corporate users do not think in terms of neat and novel, they think in terms of how much work it's going to be to keep up with the complete overhaul of a language versus moving to a language with a stable syntax once and not having to deal with it again. We will not soon rise above that kind of bad opinion. Quite. One of the nice things, so far, about working with Perl is that upgrades have been reletively simple and painless, cheap, in other words. Compared to, say, the upgrade between VB 16bit and VB 32bit, the Perl upgrades I've done have been free. The idea of changing all of my Perl scripts is *not* attractive, actually it's sort of scary. Mike . . David T. Grove Blue Square Group [EMAIL PROTECTED] [EMAIL PROTECTED]
Re: Perl, the new generation
David Grove [EMAIL PROTECTED] writes: Unless Perl 6 is capable of parsing and running that 99.9% (or higher) of Perl 5 scripts originally foretold, I foresee a far worse outcome for Perl 6 than has happened for an almost universally rejected 5.6 and 5.6.1. Most people don't adopt .0 releases. 5.6.1 was just released. Most of us who maintain large software deployments have a lag time of *at least* six months for picking up a new release of software. I know you like preaching doom and gloom, David, but you're several parsecs displaced from reality here. (And Perl 5.6.0 has been in Debian testing for a while, for that matter.) -- Russ Allbery ([EMAIL PROTECTED]) http://www.eyrie.org/~eagle/
RE: Perl, the new generation
On Thu, May 10, 2001 at 03:58:41PM -0400, David Grove wrote: it's been 13 months since 5.6 was released, and two commercial entities have so far accepted it: ActiveState and SuSE. a complete, barefaced lie. To be a lie, it must be purposeful. I am not above error, however. Who do you get your Perl from? I build my own. It's (historically speaking) the only way to get a reliable Perl on Win32, though some module still don't compile without proprietary hacks. Redhat? They ship 5.6.0 in RH7.0 Mandrake? Hrm, perl-5.600-30mdk.i586.rpm. Yep, that'd be 5.6.0 My information on this comes from discussion (asking directly) in undernet #linux. If this is in error, tell it to them. My stating this comes from actual research short of purchasing every linux on the planet just to see if they have Perl. The research took place specifically to see whether 5.6 was appropriate for PerlMagic, and it was place in 5.6 only because Win32 users thought they needed it, though several of the P5P some months ago suggested a strong warning to my users. Solaris? Talk to Alan - Perl 5.6.1 going into Solaris 9. Somebody said it was and described why. Debian? They're not commercial, but they're still a pretty big OS distro; let's have a look in the next release: (the testing distro - Debian release very infrequently.) http://packages.debian.org/testing/interpreters/perl-5.6.html shows me they're going to be shipping - oh, Perl 5.6.1. Even better. What? You mean they're actually accepting it a year and a half later in a testing version? I'm not sure you made a point here. Anywhere else? :) FreeBSD comes to mind, among others. Can we get back to the subject now? p
Re: Perl, the new generation
On Thu, May 10, 2001 at 04:41:09PM -0400, David Grove wrote: My information on this comes from discussion (asking directly) in undernet #linux. If this is in error, tell it to them. An IRC channel, in ERROR?! On Undernet no less?! THE DEUCE YOU SAY!! ;) Next thing you're going to tell me the commentary on Slashdot isn't totally impartial! Debian? They're not commercial, but they're still a pretty big OS distro; What? You mean they're actually accepting it a year and a half later in a testing version? Debian is historically slow to release 'stable' distributions. The current 'testing' branch has been going on for quite some time. FreeBSD comes to mind, among others. FreeBSD is also historically *very* slow about upgrading perl. They were one of the last hold-outs to upgrading /usr/bin/perl from perl4. Please stop. -- Michael G. Schwern [EMAIL PROTECTED]http://www.pobox.com/~schwern/ Perl6 Quality Assurance [EMAIL PROTECTED] Kwalitee Is Job One purl Hey, Schwern! THERE IS A HUGE GAZORGANSPLATTEDFARTMONGERING- LIGHTENINGBEASTASAURSOPOD BEHIND YOU! RUN, BEFORE IT GAFLUMMOXES YOUR INNARDLYBITS!
Re: Perl, the new generation
At 09:20 AM 5/10/01 -0700, I wrote: At some point, the Perl 6 cognomen will have attracted enough inertia that we couldn't reasonably change it even if we wanted to. Maybe that time has already come. Maybe not. Can't hurt to raise the question. I retract the last sentence. -- Peter Scott Pacific Systems Design Technologies http://www.perldebugged.com
Re: Perl, the new generation
On Thu, May 10, 2001 at 04:41:09PM -0400, David Grove wrote: Anywhere else? :) FreeBSD comes to mind, among others. Hm. You initially restricted your survey to commercial vendors, but now you are moving the goalposts. Can we get back to the subject now? Certainly. The subject was whether or not Perl 5.6.x has been taken up by the industry. I think we've proved that it has. Can we go back - uh, forward - to Perl 6 now? -- I think i'll take my girlfriend to vegas for a win'98 burn/upgrade -- Megahal (trained on asr), 1998-11-06
Re: Perl, the new generation
At 11:11 PM 5/10/01 +0100, Simon Cozens wrote: On Thu, May 10, 2001 at 04:41:09PM -0400, David Grove wrote: Anywhere else? :) FreeBSD comes to mind, among others. Hm. You initially restricted your survey to commercial vendors, but now you are moving the goalposts. Can we get back to the subject now? Certainly. The subject was whether or not Perl 5.6.x has been taken up by the industry. Sigh. Do I dare wade back in? But the voices in my head won't stop :-) With respect - and I do mean that - the subject as I started it was, Is Perl 6 the most appropriate title for what we discuss here and what brave people like yourself will be implementing? If it's at all possible to discuss that without devolving into tangential political debates, I'd like to do so. -- Peter Scott Pacific Systems Design Technologies http://www.perldebugged.com
RE: Perl, the new generation
On Thu, May 10, 2001 at 10:00:13PM +0100, Michael G Schwern wrote: On Thu, May 10, 2001 at 01:49:30PM -0700, Edward Peschko wrote: We need to keep syntactic compatibility, which means we need to keep the ability for perl6 to USE PERL5. I think you're in violent agreement here. This has been declared a goal of Perl 6 from almost day one. Ok, fair enough, but until just a little bit ago I was hearing stuff different from Dan. That has been changed, apparently recently If this is an actively pursued goal, I consider the issue dead, with one remark: use Perl5 or anything else that has to appear in legacy code thereby forcing hunting and changing all perl programs will likely cause it to reappear, unless we can find a way to default to both. On UN*X this is symlinks (let's not reopen it, just suffice to say that /something/ needs to prevent this requirement), and on Win32 and other systems without symlinks, a separate executable. Since the executable is so small on both (all?) systems, it's not much of an addition. Win32 already has perl5.6.0(-win32-multi-thread-morestuff).exe and perl.exe. This is an old argument, and one I don't wish to reopen, as long as /something/ along this order can provide for a painless migration. With that in place, let's change whatever syntax gets annoying. p