Web development and Perl 6

[Garry Heaton]

I'm web developer who's been usinng Perl/CGI.pm for a while before switching
to PHP.

IMHO if Perl 6 doesn't come with some kind of SDK for web developoment, ie.
at least a templating system, then it will become increasingly marginalised
in the web development world due to the proliferation of PHP within modestly
priced hosting deals.

It's not enough to say mod_perl/Embperl/Mason or whatever for the simple
reason that many new web developers and teachers use cheap hosting deals
which invariably come fully configured with PHP/MySQL and nothing more than
standard Perl/CGI.pm. These hosts won't consider running mod_perl or any
templatinng system. In fact I've come across many more expensive hosting
deals with similar constraints regarding Perl. mod_perl is too great a risk
for shared hosting environments.

Perl is in too many bits and pieces, at least where web development is
concerned. DBI, DBD::mysql, mod_perl, Mason/Embperl. That's a non-starter
with most hosting deals so you're average web developer turns to PHP which
usually comes fully optimised without the need for additional modules. I'm
talking about basic bread and butter database website work here, not LWP jobs.

Perl started out as a sysadmin tool but there's no reason to stay in that
niche now we have Perl 6 on the way. If we don't seize the chance to bundle
a proper web development SDK with Perl 6 PHP will just become the de facto
server-side scripting language for web development.

Garry Heaton

 

[Garry Heaton]

Garry Heaton ([email protected]) wrote on MMMDCXXI September MCMXCIII in
<URL:$$ I'm web developer who's been usinng Perl/CGI.pm for a while before switching
$$ to PHP.
$$
$$ IMHO if Perl 6 doesn't come with some kind of SDK for web developoment, ie.
$$ at least a templating system, then it will become increasingly marginalised
$$ in the web development world due to the proliferation of PHP within modestly
$$ priced hosting deals.
$$
$$ It's not enough to say mod_perl/Embperl/Mason or whatever for the simple
$$ reason that many new web developers and teachers use cheap hosting deals
$$ which invariably come fully configured with PHP/MySQL and nothing more than
$$ standard Perl/CGI.pm. These hosts won't consider running mod_perl or any
$$ templatinng system. In fact I've come across many more expensive hosting
$$ deals with similar constraints regarding Perl. mod_perl is too great a risk
$$ for shared hosting environments.

Oh, please. The future of Perl isn't decided by people who can't or won't
invest the money in their website that enables them to decide for themselves
which modules they have available.

The future design of Perl may not be but it usage amongst new web developers
may well be. There is currently a truckload of PHP/MySQL web development
books available. I can only think of one, Paul Dubois, for Perl and MySQL.
You have to ask why this is, or don't you care?

$$ Perl is in too many bits and pieces, at least where web development is
$$ concerned. DBI, DBD::mysql, mod_perl, Mason/Embperl. That's a non-starter
$$ with most hosting deals so you're average web developer turns to PHP which
$$ usually comes fully optimised without the need for additional modules. I'm
$$ talking about basic bread and butter database website work here, not LWP job
$$
$$ Perl started out as a sysadmin tool but there's no reason to stay in that
$$ niche now we have Perl 6 on the way. If we don't seize the chance to bundle
$$ a proper web development SDK with Perl 6 PHP will just become the de facto
$$ server-side scripting language for web development.

I don't think that would happen. However, if that's a side effect of perl6,
I'd switch my opinion from "I don't think perl6 is worth the effort" to
"I want perl6 yesterday".

I rather have that people no longer think "Perl" when they have to do
anything webbish. Or that they think "Web" if they hear "Perl".

Fine, but don't forget that Perl's popularity boom in the mid-90s was
largely due to the lack of competition in the web scripting sphere. For a
language to remain alive it has to appeal to more than the seasoned hacker.
New web developers are turning to PHP. I wish it were otherwise as I love
using Perl. However, mod_perl hosts are very thin on the ground. If new
students can get free PHP/MySQL hosting to learn web development they're
going to choose that over the complexity and relative unavailability of
mod_perl.

Garry

 

[Uri Guttman]

GH> Abigail wrote:


GH> The future design of Perl may not be but it usage amongst new web
GH> developers may well be. There is currently a truckload of
GH> PHP/MySQL web development books available. I can only think of
GH> one, Paul Dubois, for Perl and MySQL. You have to ask why this
GH> is, or don't you care?

abigail doesn't care and neither do i. perl is NOT the web and vice
versa. you need to have that concept burned into your brain. there is
far more perl code out there that has nothing to do with the web than
the other way around. you may not understand this. and much of it has
little to do with sysadmin either. perl has long grown to beyond just a
sysadmin too.

GH> Fine, but don't forget that Perl's popularity boom in the mid-90s
GH> was largely due to the lack of competition in the web scripting
GH> sphere. For a language to remain alive it has to appeal to more
GH> than the seasoned hacker. New web developers are turning to
GH> PHP. I wish it were otherwise as I love using Perl. However,
GH> mod_perl hosts are very thin on the ground. If new students can
GH> get free PHP/MySQL hosting to learn web development they're going
GH> to choose that over the complexity and relative unavailability of
GH> mod_perl.

that is utter balderdash. perl's popularity was caused by many
factors. the web was only one little but very visible part of it. and in
fact the spread of matt's scripts and other kiddie code set perl back
manyyears IMNSHO. most professionals using perl didn't need stupid
little bulletin boards and web stuff. look at cpan and see how many
modules there are that are not web related? note that the perl community
(and matt isn't included in that) never addressed dinky web scripts
until recently (the nms project). why? because it wasn't important to
anyone's job. instead real systems like mason, all the template systems,
dbi, dbi modules and such were developed. these are used by
professionals all over the place. read some of the recent perl.com
articles on web development with them. makes php look like the silly
little toy it is.


as for perl6, it aims at no particular niche such as web or anything
else. it is meant to be a full redesign of perl to reflect all the
practical knowledge about languages that has accumulated in the field
since perl5 came out. it will still be easy to write short scripts in
perl6 but it will be possible to do things almost no other language can
do and as always in less code and still be expressive. the grammar/rules
stuff alone will make perl the best language for parsing and data
munging around. all those other langs with their (not truly) PCRE
packages will have parsing envy. the OO and class stuff will leapfrog
java and c++ and others. OO will be first class and not some interesting
and simple bolted on stuff. you haven't been reading the apocalypses and
exigeses i gather. try them out before you start another misguided rant
about perl6.

uri

 

[Jürgen Exner]

I'm web developer who's been usinng Perl/CGI.pm for a while before
switching to PHP.

IMHO if Perl 6 doesn't come with some kind of SDK for web
developoment, ie. at least a templating system, then it will become
increasingly marginalised in the web development world due to the
proliferation of PHP within modestly priced hosting deals.

Sounds like a big plus to me. Maybe Perl NGs will become Perl-oriented
again?

jue

 

[Tad McClellan]

IMHO if Perl 6 doesn't come with some kind of SDK for web developoment, ie.
at least a templating system, then it will become increasingly marginalised
in the web development world due to the proliferation of PHP within modestly
priced hosting deals.

Most people in the Perl community would count that as a good thing,
so it is unlikely that there will be volunteers for such development.

Unless _you_ get some together...

It's not enough to say mod_perl/Embperl/Mason or whatever for the simple
reason that many new web developers and teachers use cheap hosting deals
which invariably come fully configured with PHP/MySQL and nothing more than
standard Perl/CGI.pm.

They get what they pay for.

PHP can have the shoe-string types, Perl can do fine without them.

Perl is in too many bits and pieces, at least where web development is
concerned. DBI, DBD::mysql, mod_perl, Mason/Embperl. That's a non-starter
with most hosting deals so you're average web developer turns to PHP which
usually comes fully optimised without the need for additional modules. I'm
talking about basic bread and butter database website work here, not LWP jobs.

If PHP is a better choice than Perl for what you need to do, then
use PHP instead of Perl. Choose the right tool for the job.

Perl and PHP can both live happy and fulfilling lives, it isn't either-or.

If we don't seize the chance to bundle
a proper web development SDK with Perl 6

Perl 6 development is done by volunteers.

Become a volunteer to do that development if you want that development done.

PHP will just become the de facto
server-side scripting language for web development.

What would be wrong with that?

Perl is useful in a boatload of other places, Perl 6 can be
used in all of those spaces.

And we'll all live happily everafter.

 

[Garry Short]

Uri Guttman wrote:

. you haven't been reading the apocalypses and
exigeses i gather. try them out before you start another misguided rant
about perl6.

uri

Hi Uri,

do you know if these are all downloadable from somewhere at the moment? I
read the early ones but am working away from home during the week at the
moment, so can't read them online. I'd love to be able to download them and
read them offline in the evenings ...

Cheers,

Garry

(not the OP)

 

[Charlton Wilbur]

The future design of Perl may not be but it usage amongst new web
developers may well be. There is currently a truckload of PHP/MySQL
web development books available. I can only think of one, Paul
Dubois, for Perl and MySQL. You have to ask why this is, or don't
you care?

As for myself, I don't much care if new web developers who want toy
scripts use Perl or not. Perl has a *far* superior set of tools for
the difficult web tasks; if PHP gets the easy ones, no skin off my
nose, because the difficult ones are more interesting.

Beyond that, MySQL is bone-simple. Even assuming that it weren't,
though, I don't need 30 good books on a topic, I need *one* good book
on a topic. The truckload of PHP/MySQL books currently exists for one
reason: publishers are fairly certain they can sell them.

Fine, but don't forget that Perl's popularity boom in the mid-90s was
largely due to the lack of competition in the web scripting sphere.

Lack of competition? I beg to differ, on two counts.

First, in the mid-90s, there were several contenders for server-side
scripting. I myself have used Tcl, C, and Scheme, as well as Perl and
server-side Java, for that purpose. From what I understand Python was
an early contender as well.

Second, Perl was well-established before the popularity boom, and was
already solidly in place as a system administration and glue language.

The reason Perl made such great inroads is that Perl has a rich set of
string manipulation tools available. So long as Perl continues to
have a rich set of tools available, it will continue to be useful. If
it becomes less relevant to clueless web newbies, so much the better.

New web developers are turning to PHP. I wish it were otherwise as I
love using Perl. However, mod_perl hosts are very thin on the
ground. If new students can get free PHP/MySQL hosting to learn web
development they're going to choose that over the complexity and
relative unavailability of mod_perl.

Let them!

It does me more benefit if the novice web users all use PHP and MySQL,
because *I* know Perl. So when someone needs a job done that's too
big for PHP and MySQL, they'll turn to Perl and Postgres or Oracle.
Benefit for me: the easy, boring stuff is being handled by eager PHP
kiddies. Benefit for me: few people know how to handle the hard,
interesting stuff. How is this a bad thing?

Charlton

 

[Gunnar Hjalmarsson]

The reason Perl made such great inroads is that Perl has a rich set
of string manipulation tools available. So long as Perl continues
to have a rich set of tools available, it will continue to be
useful. If it becomes less relevant to clueless web newbies, so
much the better.

Err.. I suppose that you started your programming career as a clueless
newbie, too. Didn't we all? (Some would say that I'm still there, btw.)

Let them!

It does me more benefit if the novice web users all use PHP and
MySQL, because *I* know Perl. So when someone needs a job done
that's too big for PHP and MySQL, they'll turn to Perl and Postgres
or Oracle. Benefit for me: the easy, boring stuff is being handled
by eager PHP kiddies. Benefit for me: few people know how to
handle the hard, interesting stuff. How is this a bad thing?

Benefit for you? Maybe. Benefit for Perl? Personally I'm not convinced.

 

[Tad McClellan]

I know many seasoned
Perl programmers who have never written a CGI script in their lives.

me <raises hand>


(except as a hobbyist)

 

[Uri Guttman]

GH> Personally I believe that any program needs a big user base, and
GH> through the web, you reach a big audience. Sometimes I feel that
GH> regulars in this group deliberately and actively discourage Perl and
GH> CGI from being used for trivial web scripting, and _that_ amazes
GH> _me_. I smell snobbery.

perl has over 1 million coders worldwide. and most don't use it for the
web. you don't seem to understand how broad and deep perl's penetration
is and how little of it is webcentric. perl would be still very popular
even if it was never used for the web. that is critical to understand.

as for snobbery, it is annoying (not snobbish) that perl has a rep for
being the web language. matt's crap archives did a lot for that. early
web cgi was almost only in perl. and still beginner cgi is usually in
perl since som many (bad) perl scripts are out there for the kiddies to
use.

but do you realize perl is used in genomics, database systems,
communications, (non-web) networking, parsing, natural language,
etc. just look at cpan (have you really ever browsed cpan's top
levels?). by far and away it has so many more non-web modules than web
stuff. and the net in general is much bigger than the web. do you ever
hear newbies (or the clueless) conflate the web and the net? i see it
all the time.

GH> Isn't CGI scripts on the web _one_ important application field for
GH> Perl? You don't need to have a "web-centric view" to acknowledge that,
GH> do you?

it is just one application. importance is irrelevant. and it is not a
large slice of perl's application world. you still don't get it. perl is
much bigger than the web. the net is much bigger than the web. the web
is just the most visible net service and the most well known by most
users. that doesn't make it important.

uri

 

[Uri Guttman]

GH> The point I was trying to make is:
GH> - Some people in this group seem to believe it is good for Perl if
GH> "newbies" opt for PHP before Perl/CGI for web scripting.
GH> - I expressed my doubts about that view.

more like it doesn't matter to perl what they choose for web stuff since
web stuff is a fraction of the entire perl world. just wait until the
newbies do something real and try to use php for genomics or parsing.

the point is not to fret over what people use for web apps. it is
nothing compared to the large picture.

uri

 

[Billyray]

Gunnar wrote:

What an amazingly web-centric view of the world! I know many seasoned
Perl programmers who have never written a CGI script in their lives.

I have discovered during the course of my life, the broader are
my experiences, the greater is my knowledge. I have also discovered
my knowledge grows at a faster rate with greater diversity of
experiences.

Blinders for mules serve a good purpose. Blinders limit a mule's
knowledge of the larger world; a mule only knows what is front
of him, keeping him relatively ignorant and more trainable but
rather narrow minded. Mules work steady and strong in front of
a plow but will clumsily break a leg if they run free.


Purl Gurl

 

[William Alexander Segraves]

Blinders for mules serve a good purpose. Blinders limit a mule's
knowledge of the larger world; a mule only knows what is front
of him, keeping him relatively ignorant and more trainable but
rather narrow minded. Mules work steady and strong in front of
a plow but will clumsily break a leg if they run free.

Hmm. For those seeking knowledge of donkeys and mules, Billy ray may not be
a good source.

For more on donkeys and mules, without blinders, from someone who is an
expert on the subject, as well as a good writer, see

http://members.aol.com/brayhaven/bj.htm

http://members.aol.com/brayhaven/bj2.htm

FYI, the word "blinder' does not appear on either page.

Bill Segraves

 

[Purl Gurl]

William Alexander Segraves ignorantly trolled:

Billyray, my main squeeze and delightful plow mule, didn't write:
(snipped)

FYI, the word "blinder' does not appear on either page.

Clearly your blinders limit if not block your mind's eye
view of my poignant point, a point highlighting this pitfall
of deliberately limiting life's experiences and lessons; my
article has nothing to do with mules nor blinders.

Trolls are not known for being exceptionally bright.

For trivia, there are no better experts on plow mules than
those who cut their teeth on plow handles under a hot
Oklahoma sun.

http://la.znet.com/~callgirl/cybefarm.html

Purl Gurl

 

[Tassilo v. Parseval]

Also sprach Uri Guttman:

GH> The point I was trying to make is:
GH> - Some people in this group seem to believe it is good for Perl if
GH> "newbies" opt for PHP before Perl/CGI for web scripting.
GH> - I expressed my doubts about that view.

more like it doesn't matter to perl what they choose for web stuff since
web stuff is a fraction of the entire perl world. just wait until the
newbies do something real and try to use php for genomics or parsing.

the point is not to fret over what people use for web apps. it is
nothing compared to the large picture.

Yet, the web is a good entrance point for a programming language. When I
started with Perl it was because I wanted to do some CGI stuff. The
interesting thing is that I soon started to use Perl for all other
things _but_ CGI (not because Perl would be unsuitable for that but
because I got bored by CGI-thingies).

For me the question remains what my primary language would be right now
hadn't Perl had the reputation to be a web-language. Perhaps I'd be
programming PHP or Java now (and be less happy).

In essence, the assumption that a person starting his life as CGI
script-kiddie will remain one till the end of his days is wrong. He
might soon switch to more interesting domains and even become a
CPAN-author or perl5/6-porter or whatever. Or become a knowledgeable
regular in clpm. Go figure.

The point of the OP was - as far as I understand it - not to alter Perl
in order to attract the kiddies. It was to provide a homogeneous Perl
environment so that webhosts are encouraged to offer a solid mod_perl
system. Would that be hard? No, not so:

perl (standard distribution or perhaps stripped down)
mod_perl
DBI
DBD::mysql
libnet
libwww
some HTML/XML modules
CGI
EmbPerl (to make it 'appear' more PHPish)

Bundle that up, provide an installation script and some documentation.
And give it a flashy name so that people start to recognize it by that
(like PerlWDK or so).

Even from a selfish point of view, that wouldn't be so bad. I might some
day be in the situation where I have use for a good webhoster. I am sure
I'll get very sick if all I have is some mod_php crap.

So all that remains to be done is take those things that already exist
and bring them together in a fairly standardized way and seamless way.

Tassilo

 

[Iain Chalmers]

The point of the OP was - as far as I understand it - not to alter Perl
in order to attract the kiddies. It was to provide a homogeneous Perl
environment so that webhosts are encouraged to offer a solid mod_perl
system. Would that be hard? No, not so:

perl (standard distribution or perhaps stripped down)
mod_perl
DBI
DBD::mysql
libnet
libwww
some HTML/XML modules
CGI
EmbPerl (to make it 'appear' more PHPish)

Bundle that up, provide an installation script and some documentation.
And give it a flashy name so that people start to recognize it by that
(like PerlWDK or so).

Seems to me that could almost all be done as a CPAN bundle...

big

 

[Garry Heaton]

Thanks Tassilo. At lesat someone has grasped my original point.

Garry Heaton

 

[Tad McClellan]

[ we lost a level of attribution somewhere here...]

Hmm.. That comment does require an explanation, Randal.

The OP seemed to be claiming that Perl 6 will fail if it is
not adopted by web developers.

Many others seem to be claiming that Perl 6 can succeed even if
it is not adopted by web developers.

With a web-centric view, then the possibility of such a failure
seems a reasonable view.

But if the preliminary assumption is wrong, then the conclusion
drawn is likely suspect.

Randal and others, are pointing out that they think the assumption
is wrong, hence so is the conclusion.

CGI scripting is (almost) all I'm using Perl for. But that is
irrelevant for this discussion.

But it is *not* irrelevant.

The topic has drifted to "web vs. traditional" development.

In "1984" fashion, we are a product of our conditioning, we "know"
only those things that are in *our* experience.

To compare two things we need experience with each of the things.

You know only web uses for Perl.

I know only non-web uses for Perl.

So what are we doing speaking up? ( rhetorical question )



Coincidentally, during a break in the middle of this thread, I
was reading the August 2003 issue of CACM [1], and found an article
related to the distinctions being drawn here:

"A Mugwump's-Eye View of Web Work" pp21-23 [2]

There's an interesting debate raging around the software
world: Is Web development the same as traditional software
project development?

with a discussion of each of the two points of view, which can
be summarized with these two quotes (I leave it to you to figure
out which quote is which side

the principles are the same, but the processes are different

we've seen all of this before

Personally I believe that any program needs a big user base, and
through the web, you reach a big audience.

Consideration of quantity withour regard to quality is a
risky approach to take.

I smell snobbery.

Me too.

If Perl 6 doesn't give me what _I_ need for web development,
then it will fail.

I expect you meant the _other_ snobbery present in this thread,
given our opposite conditionings.

Isn't CGI scripts on the web _one_ important application field for
Perl?

Cannot answer the question, as phrased.

Important to what?

Important to the folks that need to get their web app deployed? Yes.

Important to the success or failure of Perl 6? Perhaps, but there
seems to be a significant number here who'd say "No".


-----

I think what has caused this thread's size is the way the OP
presented his position.

We saw it as:

Perl 6 will fail if it doesn't do what "web developers" need.

which prompted followups pointing out that that segment is not
as large a proportion as the OP seems to think.

There would have been much less contention if he'd instead
said something like:

I really need this in my work, can Perl 6 be made to give it to me?


The Perl developers are very interested in helping Perl6ers
"get their work done". That is what Perl is, and always has been,
all about.

An appeal to that interest would have a much better chance of
leading to action rather than merely discussion.

I'm sure folks would _like_ for Perl 6 to be a reasonable choice
of language for CGI programming.

But there is an important distinction between "it would be nice"
and "it is essential".




[1] "Communications of the ACM"

[2] mugwump: A person who acts independently or remains neutral

 

[William Alexander Segraves]

If the mailing address for ''Billyray'' is accurate, she
may not be a good source for those seeking knowledge on
any subject, though asses might just be up her alley.

Time to adjust your scorefiles, kids!

Apologies for adding to off-topic noise, folks.

Clearly, Randal needs no defense.

Bill Segraves

 

[Charlton Wilbur]

Err.. I suppose that you started your programming career as a clueless
newbie, too. Didn't we all? (Some would say that I'm still there, btw.)

Yes; but at that time there was no web in common usage, and I wasn't
arguing to seasoned programmers that all programming languages should
be completely accessible to clueless web newbies, and neither was I
under the delusion that a book on a popular language would make me
anything BUT a clueless newbie with a book in his hand. This,
perhaps, makes me substantially less clueless than the average
clueless web newbie.

Charlton