Re: Web Module (Was: Perl6 new features)

[Peter Scott]

On Mon, 25 Jun 2007 20:30:36 +0200, Juerd Waalboer wrote:
> A useful, easily installable library bundle does not have to be in the
> core distribution.
> 
> Debian already has its own split between perl and perl-modules. This is
> a great scheme that allows Debian to use Perl in the base system,
> without requiring the loads of unused modules that usually come with it.
> If you need those modules, though, you can easily install them.

That is such a specialized application that it would make sense to release
two distributions of Perl:

Perl 6
Perl 6 for OS and embedded systems integrators

-- 
Peter Scott
http://www.perlmedic.com/
http://www.perldebugged.com/

Re: Web Module (Was: Perl6 new features)

[chromatic]

On Monday 25 June 2007 00:57:18 Hakim Cassimally wrote:

> Releasing a language without a useful, easily installable library bundle
> could quite reasonably be construed as a stupid business practice.

Of course.  Yet some dozen years later, the argument for keeping interfaces 
such as File::Find (because sysadmins should learn all about closures and 
callbacks but can only understand global variables) and code such as CGI.pm 
(it was SelfLoader before SelfLoader was cool and *that* was a while ago) in 
the Perl 5 core has absolutely nothing to do with quality, ease of use, or 
suitability for the problem domain and everything to do with incidentally 
having been first and, thus, immediate evolutionary dead ends.

-- c

Re: Web Module (Was: Perl6 new features)

[Juerd Waalboer]

Hakim Cassimally skribis 2007-06-25  9:57 (+0200):
> Releasing a language without a useful, easily installable library bundle
> could quite reasonably be construed as a stupid business practice.

A useful, easily installable library bundle does not have to be in the
core distribution.

Debian already has its own split between perl and perl-modules. This is
a great scheme that allows Debian to use Perl in the base system,
without requiring the loads of unused modules that usually come with it.
If you need those modules, though, you can easily install them.

It would be great if Perl had this by default, because it would make it
easier for vendors to choose to use Perl in their base system. It would
also make Perl a more attractive choice for embedded systems.
-- 
korajn salutojn,

  juerd waalboer:  perl hacker  <[EMAIL PROTECTED]>  
  convolution: ict solutions and consultancy <[EMAIL PROTECTED]>

Re: Web Module (Was: Perl6 new features)

[Paul Hodges]

How about a Bundle::Common?
Streamline both the core and the inclusion of the most commonly used
modules? The core does include the CPAN module, right?

Personally, I *prefer* grabbing what I need piecemeal, but I understand
making it easy if possible

--- Luke Palmer <[EMAIL PROTECTED]> wrote:

> On 6/25/07, Peter Scott <[EMAIL PROTECTED]> wrote:
> > I will just voice my support for putting best-of-breed modules for
> very
> > common tasks (CGI, DBI for sure) in the core.
> 
> Of course, then you get the disadvantage that most users will see new
> versions of those modules as often (or seldom, as it were) as perl
> itself.
> 
> Luke
> 


===
Hodges' Rule of Thumb: Don't expect reasonable behavior from anything with a 
thumb.


   

Choose the right car based on your needs.  Check out Yahoo! Autos new Car 
Finder tool.
http://autos.yahoo.com/carfinder/

Re: Web Module (Was: Perl6 new features)

[Luke Palmer]

On 6/25/07, Peter Scott <[EMAIL PROTECTED]> wrote:

I will just voice my support for putting best-of-breed modules for very
common tasks (CGI, DBI for sure) in the core.


Of course, then you get the disadvantage that most users will see new
versions of those modules as often (or seldom, as it were) as perl
itself.

Luke

Re: Web Module (Was: Perl6 new features)

[Peter Scott]

On Mon, 25 Jun 2007 09:57:18 +0200, Hakim Cassimally wrote:
> On 23/06/07, chromatic <[EMAIL PROTECTED]> wrote:
>> On Friday 22 June 2007 11:07:35 Chas Owens wrote:
>>
>> > Please, god, no. Or at least make two distributions: Bare Perl 6 and
>> > Perl 6. Many companies have a "Only Core Perl" policy. They refuse
>> > to install CPAN modules because "We don't trust them".
>>
>> I think of this the same way I think of "Do not drink even if you mix with
>> beer" labels on bleach bottles.  Stupid people often remove their genes from
>> the pool.  We should not discourage stupid business practices from doing the
>> same.
> 
> Releasing a language without a useful, easily installable library bundle
> could quite reasonably be construed as a stupid business practice.

I will just voice my support for putting best-of-breed modules for very
common tasks (CGI, DBI for sure) in the core.  It may not be elegant, but
it is practical... and Perl has weighed in on that side of the dilemma
before.  I think part of Perl 4's success was the "one stop shopping"
availability of numerous useful functions that would otherwise have
required linking in a library, e.g. dbmopen, crypt, shmget, gethostbyaddr,
math functions, etc.  I think the synergistic effect of the combination of
all these functions being so readily available helped Perl take off. 
Please continue the tradition of tending to err on the side of "practical"
over "elegant".  It may not seem rational that avoiding the need to type
"cpan Foo::Bar" makes a huge difference, but I believe it does, for
certain modules.

-- 
Peter Scott
http://www.perlmedic.com/
http://www.perldebugged.com/

Re: Web Module (Was: Perl6 new features)

[Hakim Cassimally]

On 23/06/07, chromatic <[EMAIL PROTECTED]> wrote:

On Friday 22 June 2007 11:07:35 Chas Owens wrote:

> Please, god, no. Or at least make two distributions: Bare Perl 6 and
> Perl 6. Many companies have a "Only Core Perl" policy. They refuse
> to install CPAN modules because "We don't trust them".

I think of this the same way I think of "Do not drink even if you mix with
beer" labels on bleach bottles.  Stupid people often remove their genes from
the pool.  We should not discourage stupid business practices from doing the
same.


Releasing a language without a useful, easily installable library bundle
could quite reasonably be construed as a stupid business practice.

I do hope that Perl6 doesn't do this, but a) I don't think it's very likely,
and b) some organization or company will rapidly release a useful
Bundle:: (or Task:: or whatever it's called in Perl6) to fill in the gap.
--
osfa

Re: Web Module (Was: Perl6 new features)

[cdumont]

Getting back to the Web Module,
the following could help a lot
if there were in the core or as a 'core' downloadable bundle :

- DBI
- ORM system
- Caching system
- Sessions (server side cookies)
- XML parser
- JSON parser

I guess that it will be fair enough for actual web apps.

templating system that could be hooked or using flex points,
classes related to CRUD,scaffolding,input filtering, unit testing could
be a plus
but could sound more like a framework than a programming language^^;

I've certainly forgotten many things but that's what came up at the moment.

Re: Web Module (Was: Perl6 new features)

[DPU-ibrown]

The fact that you'll be able to do that in Perl6 excites me.  One of the
things I use with the existing Perl5 unfortunately at times, is
commercial software which compiles Perl code into various Microsoft
formats: services, system tray icons, dll's and executables.  

That proved to be extremely valuable to me when I was thrown the
responsibility of designing implementing and maintaining an Active
Directory environment.  I've been writing Perl for years, and truly did
not want to learn to code MS's way.  I liked Perl.

I think that it would be truly a valuable thing to have the ability to
distribute Perl as executable binaries, completely stand alone and
independance from the need of module installation on the client side.  

Especially with the way things are going in the market place, if we can
truly get cross-platform compiles that would be in the words of a tru
comedian, "Dyno--mite".

Perl saves me from proprietaries *sic* on a daily basis.  Its easy for
me to pay for a commercial product to compile Perl5 into "executable
modules" i release for system changes via Group Policies etc.  I just
wish I didn't have to pay for that commercial product to help me love
Perl.

-I
On Fri, 2007-06-22 at 18:36 -0400, Chas Owens wrote:
> On 6/22/07, chromatic <[EMAIL PROTECTED]> wrote:
> > On Friday 22 June 2007 11:07:35 Chas Owens wrote:
> >
> > > Please, god, no. Or at least make two distributions: Bare Perl 6 and
> > > Perl 6. Many companies have a "Only Core Perl" policy. They refuse
> > > to install CPAN modules because "We don't trust them".
> >
> > I think of this the same way I think of "Do not drink even if you mix with
> > beer" labels on bleach bottles.  Stupid people often remove their genes from
> > the pool.  We should not discourage stupid business practices from doing the
> > same.
> >
> > -- c
> >
> 
> I still need to earn a paycheck, and I would rather do it with Perl
> than with Java.  Fighting my customers stupidity, instead of working
> around it, is a good way to stop earning my paycheck. But, as Jerry
> pointed out, it will be a lot easier to hide module usage in Perl 6.
> Get a code review sign off from people who aren't idiots, compile it
> all to bytecode, and hand off the binary file to the schmucks who are
> the problem.

Re: Web Module (Was: Perl6 new features)

[Chas Owens]

On 6/22/07, chromatic <[EMAIL PROTECTED]> wrote:

On Friday 22 June 2007 11:07:35 Chas Owens wrote:

> Please, god, no. Or at least make two distributions: Bare Perl 6 and
> Perl 6. Many companies have a "Only Core Perl" policy. They refuse
> to install CPAN modules because "We don't trust them".

I think of this the same way I think of "Do not drink even if you mix with
beer" labels on bleach bottles.  Stupid people often remove their genes from
the pool.  We should not discourage stupid business practices from doing the
same.

-- c



I still need to earn a paycheck, and I would rather do it with Perl
than with Java.  Fighting my customers stupidity, instead of working
around it, is a good way to stop earning my paycheck. But, as Jerry
pointed out, it will be a lot easier to hide module usage in Perl 6.
Get a code review sign off from people who aren't idiots, compile it
all to bytecode, and hand off the binary file to the schmucks who are
the problem.

Re: Web Module (Was: Perl6 new features)

[chromatic]

On Friday 22 June 2007 11:07:35 Chas Owens wrote:

> Please, god, no.  Or at least make two distributions: Bare Perl 6 and
> Perl 6.  Many companies have a "Only Core Perl" policy.  They refuse
> to install CPAN modules because "We don't trust them".

I think of this the same way I think of "Do not drink even if you mix with 
beer" labels on bleach bottles.  Stupid people often remove their genes from 
the pool.  We should not discourage stupid business practices from doing the 
same.

-- c

Re: Web Module (Was: Perl6 new features)

[Chas Owens]

On 6/22/07, jerry gay <[EMAIL PROTECTED]> wrote:

On 6/22/07, Chas Owens <[EMAIL PROTECTED]> wrote:
> Most of the time the policy is enacted by lower-case-l lazy sysadmins
> who can't be bothered to type
>
> perl -MCPAN -e install Foo::Bar
>
> My normal route around them is to install the module into the home
> directory of the user who is going to run the script, but I have had
> difficulty with this before when it comes time to move to production:
> "Where is the code review for that code?".  My answer of "where is the
> code review for that (often open source) database install you just
> did?" doesn't tend to hold the weight I wish it did.  For some reason
> binary blobs make some types of sysadmins feel all fuzzy and warm
> inside.
>
so use the parrot back end and compile all the modules to bytecode.
oh, and you can merge the foreign module bytecode with the bytecode
for your application, so it's all one big happy binary file.

in fact, parrot will even provide a way to compile bytecode to a
native executable which contains parrot itself. there, now you've got
a proper binary with *zero* external requirements in the production
environment--it doesn't even need to have parrot installed.

at that point, i'd be surprised if the release engineers or sysadmins
even notice.
~jerry



Good point.  I am still to stuck in the Perl 5 mind set.

Re: Web Module (Was: Perl6 new features)

[jerry gay]

On 6/22/07, Chas Owens <[EMAIL PROTECTED]> wrote:

Most of the time the policy is enacted by lower-case-l lazy sysadmins
who can't be bothered to type

perl -MCPAN -e install Foo::Bar

My normal route around them is to install the module into the home
directory of the user who is going to run the script, but I have had
difficulty with this before when it comes time to move to production:
"Where is the code review for that code?".  My answer of "where is the
code review for that (often open source) database install you just
did?" doesn't tend to hold the weight I wish it did.  For some reason
binary blobs make some types of sysadmins feel all fuzzy and warm
inside.


so use the parrot back end and compile all the modules to bytecode.
oh, and you can merge the foreign module bytecode with the bytecode
for your application, so it's all one big happy binary file.

in fact, parrot will even provide a way to compile bytecode to a
native executable which contains parrot itself. there, now you've got
a proper binary with *zero* external requirements in the production
environment--it doesn't even need to have parrot installed.

at that point, i'd be surprised if the release engineers or sysadmins
even notice.
~jerry

Re: Web Module (Was: Perl6 new features)

[Chas Owens]

On 6/22/07, Daniel Hulme <[EMAIL PROTECTED]> wrote:

On Fri, Jun 22, 2007 at 02:07:35PM -0400, Chas Owens wrote:
> On 6/22/07, chromatic <[EMAIL PROTECTED]> wrote:
> >I also like to proceed from the assumption that the only core modules
> >should be those required to install other modules.

> Please, god, no.  Or at least make two distributions: Bare Perl 6 and
> Perl 6.  Many companies have a "Only Core Perl" policy.  They refuse
> to install CPAN modules because "We don't trust them".  Yeah, it is
> stupid, but as a contractor I have limited ability to fight their
> ignorance.
Sounds like a bare Perl 6 distribution might be just what you need to
get them to weaken their restriction to "only modules with a good
cpanrating" or, best of all, "evaluate each module according to its
merits." Well, actually, I'm not sure that last one would be best, as
getting each module evaluating would almost inevitably entail getting
five or six high-up people together, none of whom have any interest in
you getting your job done, and who probably hate each others' guts, and
getting them to commit to some form of responsibility-generating paper-
trail.

Call me cynical if you like, but I prefer to call it, "experienced."


Most of the time the policy is enacted by lower-case-l lazy sysadmins
who can't be bothered to type

perl -MCPAN -e install Foo::Bar

My normal route around them is to install the module into the home
directory of the user who is going to run the script, but I have had
difficulty with this before when it comes time to move to production:
"Where is the code review for that code?".  My answer of "where is the
code review for that (often open source) database install you just
did?" doesn't tend to hold the weight I wish it did.  For some reason
binary blobs make some types of sysadmins feel all fuzzy and warm
inside.

As to whether a bare bones Perl distribution would help or hinder the
process of getting people to accept modules, well, all I can say is
that it is hard enough to get Perl into some companies.  If I had to
tell them "Oh, and we have to install all of these modules before we
can start coding." their response would likely be "Java already can do
that.  Use it instead.".  Yes, I have been told that before (related
to some XML stuff).

Re: Web Module (Was: Perl6 new features)

[Daniel Hulme]

On Fri, Jun 22, 2007 at 02:07:35PM -0400, Chas Owens wrote:
> On 6/22/07, chromatic <[EMAIL PROTECTED]> wrote:
> >I also like to proceed from the assumption that the only core modules 
> >should be those required to install other modules.

> Please, god, no.  Or at least make two distributions: Bare Perl 6 and
> Perl 6.  Many companies have a "Only Core Perl" policy.  They refuse
> to install CPAN modules because "We don't trust them".  Yeah, it is
> stupid, but as a contractor I have limited ability to fight their
> ignorance.
Sounds like a bare Perl 6 distribution might be just what you need to
get them to weaken their restriction to "only modules with a good
cpanrating" or, best of all, "evaluate each module according to its
merits." Well, actually, I'm not sure that last one would be best, as
getting each module evaluating would almost inevitably entail getting
five or six high-up people together, none of whom have any interest in
you getting your job done, and who probably hate each others' guts, and
getting them to commit to some form of responsibility-generating paper-
trail.

Call me cynical if you like, but I prefer to call it, "experienced."

-- 
"It must be accepted as a principle that the rifle,  effective as it is,
cannot  replace  the effect  produced  by the  speed of  the horse,  the
magnetism of the charge, and the terror of cold steel." 
  -- British Cavalry training manual, 1907 ::: http://surreal.istic.org/


signature.asc
Description: Digital signature

Re: Web Module (Was: Perl6 new features)

[Chas Owens]

On 6/22/07, chromatic <[EMAIL PROTECTED]> wrote:

On Thursday 21 June 2007 15:23:38 Smylers wrote:

> Has Larry yet decreed whether Web will be bundled with Perl 6?

I also like to proceed from the assumption that the only core modules should
be those required to install other modules.

-- c



Please, god, no.  Or at least make two distributions: Bare Perl 6 and
Perl 6.  Many companies have a "Only Core Perl" policy.  They refuse
to install CPAN modules because "We don't trust them".  Yeah, it is
stupid, but as a contractor I have limited ability to fight their
ignorance.

Re: Web Module (Was: Perl6 new features)

[chromatic]

On Thursday 21 June 2007 15:23:38 Smylers wrote:

> Has Larry yet decreed whether Web will be bundled with Perl 6?

I also like to proceed from the assumption that the only core modules should 
be those required to install other modules.

-- c

Re: Web Module (Was: Perl6 new features)

[Smylers]

Darren Duncan writes:

> At 11:23 PM +0100 6/21/07, Smylers wrote:
> 
> > Has Larry yet decreed whether Web will be bundled with Perl 6?
> 
> I believe that what Larry has said is that there are no official core
> modules, ... So ... not something to worry about now.

Thanks for that, Darren.

Smylers

Re: Web Module (Was: Perl6 new features)

[Smylers]

Juerd Waalboer writes:

> Smylers skribis 2007-06-21 23:23 (+0100):
> 
> > Of course.  But there's a big difference between the attitude of
> > 'let's do the best we can right now' and 'this is our one chance to
> > do this right'.
> 
> I think that for some things, mainly for setting community standards
> (Web, POD, ...), this is our one chance to get it right, for the next
> two decades.

Fair enough.

Obviously it would be great if that happened, and by the time the code
part of Perl 6 is releasable we have all those other things.  But if we
get to the state where core Perl 6 is ready to unleash on the world as a
stable release, but it's still the case that:

* The OO documentation, while readable and complete, if given a little
  more structure would be better for computer parsing.

* We haven't finalized the Web module.

* There isn't a fully working Perl 6 implementation of 'Duke Nukem
  Forever' in the examples/ subdirectory.

... then I don't think it would be worth holding up Perl 6.0.0.'s
release to wait for any of those things to happen.

Smylers

Re: Web Module (Was: Perl6 new features)

[Darren Duncan]

At 11:23 PM +0100 6/21/07, Smylers wrote:

Has Larry yet decreed whether Web will be bundled with Perl 6?


I believe that what Larry has said is that there are no official core 
modules, but that people who distribute Perl can bundle whatever they 
want, and that such choice was to be encouraged.


More generally, I would expect that the only ext/Foo/ that would 
surely be bundled with Perl are those that implement Synopsis-defined 
features, eg that Math one, and those are very few.


And by the time this is something to worry about, ext/ probably would 
have gone away, with required modules being somewhere else, such as 
the Prelude is, or distributed separately on Perl 6 CPAN.


As for whether ext/HTTP/ and ext/Web/ are typically bundled, I 
suspect that's just something that will happen defacto later on, and 
Larry won't decree the matter.


So ... not something to worry about now.

-- Darren Duncan

Re: Web Module (Was: Perl6 new features)

[Juerd Waalboer]

Smylers skribis 2007-06-21 23:23 (+0100):
> Of course.  But there's a big difference between the attitude of 'let's
> do the best we can right now' and 'this is our one chance to do this
> right'.

I think that for some things, mainly for setting community standards
(Web, POD, ...), this is our one chance to get it right, for the next
two decades.  

This said, I don't think Web should be core. It should be the *first*
web toolkit for Perl 6, though.
-- 
korajn salutojn,

  juerd waalboer:  perl hacker  <[EMAIL PROTECTED]>  
  convolution: ict solutions and consultancy <[EMAIL PROTECTED]>

Re: Web Module (Was: Perl6 new features)

[Smylers]

Moritz Lenz writes:

> Smylers wrote:
> 
> > Moritz Lenz writes:
> > 
> > > Web is hopefully "CGI done right"
> > 
> > ... why are we hoping that it will be "done right"?  
> 
> Because we hope we learned from the past. There are several other
> modules that fullfill most of CGI's tasks, some of them do most of it
> better. Reimplementing one of them would already make Web a "better
> CGI" ;-)

Indeed -- making a better CGI is a great idea.

But the fact that such innovations came along later supports my point.
I suspect that some of these got less mindshare than they otherwise
would have done (or were seen as inferior to CGI) because:

* The CGI module was core.

* By being called simply CGI, the CGI module gets some kind of
  superiority over all the other modules which have to be
  CGI::Something.

Has Larry yet decreed whether Web will be bundled with Perl 6?

> > Given the history of things like this it strikes me as unlikely that
> > now is the particular moment where we suddenly manage to create a
> > perfect library, and as such this would be hoping against the light
> > of available evidence!
> 
> You're right, but we should ship things as best as we can, so we try
> ;-)

Of course.  But there's a big difference between the attitude of 'let's
do the best we can right now' and 'this is our one chance to do this
right'.

> > It seems entirely possible that during Perl 6's life somebody,
> > possibly somebody who at the moment hasn't even heard of Perl 6,
> > will create a better web module.  It would be good if at that point
> > it becomes straightforward for it to get acceptance and people to
> > adopt it.
> 
> Right, but that's no reason not to try hard on your own.

It isn't.  But it is a reason to anticipate the existence of future
developments, and try to be careful not to do anything which makes life
harder for them.

Smylers

Re: Web Module (Was: Perl6 new features)

[Moritz Lenz]

Smylers wrote:
> Moritz Lenz writes:
> 
>> You could help by contributing some suggestions to what the new "Web"
>> module should be able to do, and how so. Web is hopefully "CGI done
>> right", and still in its early planning stage.
> 
> Web module?  This is the first I've heard of it.  Where is it being
> planned, if not on this list?
> 
> Also, why are we hoping that it will be "done right"?  

Because we hope we learned from the past. There are several other
modules that fullfill most of CGI's tasks, some of them do most of it
better. Reimplementing one of them would already make Web a "better CGI" ;-)

> Given the history
> of things like this it strikes me as unlikely that now is the particular
> moment where we suddenly manage to create a perfect library, and as such
> this would be hoping against the light of available evidence!

You're right, but we should ship things as best as we can, so we try ;-)

> It seems entirely possible that during Perl 6's life somebody, possibly
> somebody who at the moment hasn't even heard of Perl 6, will create a
> better web module.  It would be good if at that point it becomes
> straightforward for it to get acceptance and people to adopt it.

Right, but that's no reason not to try hard on your own.

Moritz

-- 
Moritz Lenz
http://moritz.faui2k3.org/ |  http://perl-6.de/



signature.asc
Description: OpenPGP digital signature

Re: Web Module (Was: Perl6 new features)

[Smylers]

Darren Duncan writes:

> At 6:37 PM +0100 6/21/07, Smylers wrote:
>
> > Web module?  This is the first I've heard of it.  Where is it being
> > planned, if not on this list?
> 
> It was being discussed on the perl6-users list, last year.

Thanks.

Smylers

Re: Web Module (Was: Perl6 new features)

[Darren Duncan]

At 6:37 PM +0100 6/21/07, Smylers wrote:

Moritz Lenz writes:
 > You could help by contributing some suggestions to what the new "Web"

 module should be able to do, and how so. Web is hopefully "CGI done
 right", and still in its early planning stage.


Web module?  This is the first I've heard of it.  Where is it being
planned, if not on this list?


It was being discussed on the perl6-users list, last year.  And see 
ext/Web/ in the Pugs repo.  Actually, see ext/HTTP/ first, since the 
2 come in a pair, as a result of said discussion, and have been in 
the Pugs repo since Feb 17th.  Actually, see 
http://svn.pugscode.org/pugs/ext/HTTP/docs/ first, which is a copy of 
the important emails by Juerd which the HTTP+Web were started from. 
-- Darren Duncan

Re: Web Module (Was: Perl6 new features)

[Smylers]

Moritz Lenz writes:

> You could help by contributing some suggestions to what the new "Web"
> module should be able to do, and how so. Web is hopefully "CGI done
> right", and still in its early planning stage.

Web module?  This is the first I've heard of it.  Where is it being
planned, if not on this list?

Also, why are we hoping that it will be "done right"?  Given the history
of things like this it strikes me as unlikely that now is the particular
moment where we suddenly manage to create a perfect library, and as such
this would be hoping against the light of available evidence!

It seems entirely possible that during Perl 6's life somebody, possibly
somebody who at the moment hasn't even heard of Perl 6, will create a
better web module.  It would be good if at that point it becomes
straightforward for it to get acceptance and people to adopt it.

Smylers