discuss: Resources for TLDP (was: Re: TLDP and the COmmercial World)
Subject:
Re: Resources for TLDP (was: Re: TLDP and the COmmercial World)
From:
Stein Gjoen ####@####.####
Date:
4 Mar 2005 15:57:28 -0000
Message-Id: <42288673.6080809@mail.nyx.net>
####@####.#### wrote:
>
>> Have a look at my prototype page at
>> http://www.nyx.net/~sgjoen/tarpl.html
>> One of the two generators there is a mock-up for defining what
>> parts of the collection you want, how you want it and then start
>> downloading it. Making it work for real would be nice.
>
>
> I've had a look at The LDP Collection Generator and the Table of HOWTOs.
>
> GENERATOR
>
> I think the Generator would be relatively straightforward provided all
> the tools required were on the server. Do you see the end product as a
> RPM/DEB/TGZ, or simply a compressed archive of some kind?
The Collection Generator is a mockup for illustrating the general
points so I guess I should update it soon with more details and
specifics. I see end product as two cases:
- a compressed archive that is defined, downloaded
and then unpacked in a private directory for further
modifications and customisation for, say, inclusion
on a magazine CD-ROM
- a package (RPM, DEB, TGZ) for installing into a new
or existing distribution
These archives should hopefully be designed to comply with the
LDP File Hierarchy Standard
http://www.nyx.net/~sgjoen/LFH.txt
> Do you have any preferences for language (Perl/Python/PHP)? Perl/Python
> would probably be CGI, PHP would probably be an Apache module.
I have no language preferences. I am however very glad to hear
you consider this straightforward.
> I think we could keep the current HTML page and send the form data to a
> separate script.
A little update wold probably be in order, I hope others can take a
look to see if anything is missing. Early version will probably not
feature all possibilities (such as RFC and TR) but I believe it is
important to start early and start simple and then add features in
a controlled manner over time.
For instance it would be useful to be able to download a patch
from an earlier version, perhaps based on weekly baselines. This
has many benefits:
- easier selection
- smaller download, simpler installation/update
- significantly less bandwidth requirements
> Generation could all happen dynamically on-demand, or maybe one set of
> downloads could be generated daily/hourly. The best way to go would
> depend on how many downloads occurred daily (the more downloads, the
> more sense static pre-generated files would make).
>
> TABLE
>
> I'm curious how this table is generated. I can imagine XSL, SQL, perl
> script and manual methods.
I guess this is the The Document Index and Download Page:
http://www.nyx.net/~sgjoen/doc_index.html
This version is generated by hand, another mockup that underneath a
nice interface is in reality ugly as sin. I had to make a version
that offended my senses a little less, using simple JavaScript:
http://www.nyx.net/~sgjoen/doc_index_v2.html
Again it was manual but also less painful. It should be simpler
to automate. Equally important it is requires far less
bandwidth: For 3 entries we have 4.8 KB vs. 6.3 KB.
Also this one should be updated with a number of things:
- feedback/rating method: no agreement on this yet
- keywords: not implemented yet
- categories: should be added, a HOWTO can have up to 3 categories
- formats: we need an agreement on what formats we wish to provide
With more discussions I am sure others will add more. Still I
hope we can start simple and continue from there.
Regards,
Stein Gjoen
