--Apple-Mail-14-857294617
Content-Transfer-Encoding: 7bit
Content-Type: text/plain;
	charset-ASCII;
	format­´wed


On Sunday, January 26, 2003, at 02:37  AM, David Garamond wrote:

> i'm not saying that we should copycat CPAN. but when people come to 
> this group from time to time asking about a CPAN-like beast, that 
> ought to make us think about what features CPAN has that RAA currently 
> doesn't. and can we make a better cpan than CPAN (just like ruby is in 
> many aspects a better Perl than perl).

I have to say that, for me as someone who remembers perl 4 and is just 
starting Ruby, the big thing missing to make RAA an acceptable 
alternative to CPAN is not a website, but the CPAN module. I NEVER use 
the perl CPAN ftp-site/web-page - I always access it through the 
module. If someone built something equivalent to CPAN.pm, hopefully 
with a less kludgy interface, I think it would go a long way.

Also, a not on categories: when I use the CPAN shell, I don't care 
about categories, because my workflow is:

 > i /FTP/
... blah ...
... blah ...
 > readme DEC::FTP::OverDecnet::Perl
... blah ...
... blah ...
 > install Net::FTP::OverDecnet::Perl
... blah ...
# OOPS
 > install Dec::FTP::OverDecnet::Perl
... random compiler errors ...
... random attempts to install the latest perl ...
... random additional modules required ...
 > force install Dec::FTP::OverDecnet::Perl
... blah ...
... blah ...
 > quit

The point is that I NEVER look at what category a module's in because I 
always use the search function.  I suspect that most people do the same 
- CPAN's just too big and has too much junk to use it otherwise.

If Ruby wants to build a better CPAN, I would suggest a few things:

1 - A better interface. The CPAN shell is kludgy at best.
2 - Better dependency handling and management. With a good interface, I 
should be able to choose where to get the modules I need.
3 - Easy ways to get data into the directory.

I think that sourceforge might be a better model than CPAN - CPAN 
followed the lead of CTAN, which was built in the days when FTP was 
king. We have the ability, today, to build something more interesting, 
more web-based, and more database-oriented. I think we should focus on 
that rather than on category listings - especially since the categories 
always seem to be more or less inappropriate anyway.

Patrick


--Apple-Mail-14-857294617
Content-Transfer-Encoding: 7bit
Content-Type: text/enriched;
	charset-ASCII



On Sunday, January 26, 2003, at 02:37  AM, David Garamond wrote:


<excerpt><fixed>i'm not saying that we should copycat CPAN. but when
people come to this group from time to time asking about a CPAN-like
beast, that ought to make us think about what features CPAN has that
RAA currently doesn't. and can we make a better cpan than CPAN (just
like ruby is in many aspects a better Perl than perl).</fixed>

</excerpt>

I have to say that, for me as someone who remembers perl 4 and is just
starting Ruby, the big thing missing to make RAA an acceptable
alternative to CPAN is not a website, but the CPAN module. I NEVER use
the perl CPAN ftp-site/web-page - I always access it through the
module. If someone built something equivalent to CPAN.pm, hopefully
with a less kludgy interface, I think it would go a long way. 


Also, a not on categories: when I use the CPAN shell, I don't care
about categories, because my workflow is:


> i /FTP/

... blah ...

... blah ...

> readme DEC::FTP::OverDecnet::Perl

... blah ...

... blah ...

> install Net::FTP::OverDecnet::Perl

... blah ...

# OOPS

> install Dec::FTP::OverDecnet::Perl

... random compiler errors ...

... random attempts to install the latest perl ...

... random additional modules required ...

> force install Dec::FTP::OverDecnet::Perl

... blah ...

... blah ...

> quit


The point is that I NEVER look at what category a module's in because
I always use the search function.  I suspect that most people do the
same - CPAN's just too big and has too much junk to use it otherwise. 


If Ruby wants to build a better CPAN, I would suggest a few things:


1 - A better interface. The CPAN shell is kludgy at best.

2 - Better dependency handling and management. With a good interface,
I should be able to choose where to get the modules I need.

3 - Easy ways to get data into the directory.


I think that sourceforge might be a better model than CPAN - CPAN
followed the lead of CTAN, which was built in the days when FTP was
king. We have the ability, today, to build something more interesting,
more web-based, and more database-oriented. I think we should focus on
that rather than on category listings - especially since the
categories always seem to be more or less inappropriate anyway.


Patrick



--Apple-Mail-14-857294617--