Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 343457 - Gentoo.org - redesign
Summary: Gentoo.org - redesign
Status: RESOLVED CANTFIX
Alias: None
Product: Websites
Classification: Unclassified
Component: Other (show other bugs)
Hardware: All Linux
: High normal (vote)
Assignee: Gentoo Infrastructure
URL: http://gentoo.org
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-10-31 04:52 UTC by denysonique
Modified: 2010-11-02 02:55 UTC (History)
2 users (show)

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description denysonique 2010-10-31 04:52:13 UTC
currently gentoo.org looks very ugly the website could be redesigned to look more clean and be more readable.

this idea to fill this bug report sparked on #gentoo-chat while ormaj, me and tatsh and others where chatting. we beleive that the graphic look of gentoo.org could be firmly improved.




Reproducible: Always

Steps to Reproduce:
1. visit http://gentoo.org
2. notice how ugly it is
Comment 1 denysonique 2010-10-31 05:00:25 UTC
http://denysonique.tk/gentoo.org.png (it was not possible to upload, so I am posting a link)

This is a screenshot of a small example of how for example gentoo.org could be improved, ormajj and I, came up with this little enhancement. this is only an example of what could be done
Comment 2 denysonique 2010-10-31 05:06:18 UTC
sorry the link is http://denysonique.tk/gentoo/gentoo.org.png
Comment 3 Andrew Udvare 2010-10-31 05:12:15 UTC
That link is dead.

I have one suggestion, move to Drupal 7! It's almost stable and very easy to fix, patch, make look pretty, etc etc.

A lot of FOSS projects are using this Drupal installation profiles to get up and running really quickly:  http://drupal.org/project/Installation+profiles .

Example: http://www.clementine-player.org/

A site as technical as Gentoo.org doesn't really need a crazy eye-candy theme. That's what I do not want to see. However, when I say this, I don't mean go GNU style with only what you need (<html>,<head>,<body>,<h1> and maybe <h2>,<p>, and minimal CSS). It's not a bad idea but most of us have good browsers now. Still have to make sure the site is accessible by text browsers so those who are installing who don't have a GUI yet but have Internet may access. So using semantic HTML (as GNU does) is very important.
Comment 4 Alex Legler (RETIRED) archtester Gentoo Infrastructure gentoo-dev Security 2010-10-31 08:33:59 UTC
(In reply to comment #0)
> currently gentoo.org looks very ugly the website could be redesigned to look
> more clean and be more readable.
> 
> this idea to fill this bug report sparked on #gentoo-chat while ormaj, me and
> tatsh and others where chatting. we beleive that the graphic look of gentoo.org
> could be firmly improved.

It's a known fact that our website needs a relaunch.
Sadly, just a few rounded corners and shadows won't fix that.

(In reply to comment #3)
> That link is dead.
> 
> I have one suggestion, move to Drupal 7! It's almost stable and very easy to
> fix, patch, make look pretty, etc etc.
> 

That's not going to happen (imo).
Comparing the clementine website you showed as an example, I count <10 actual pages. Our website contains hundreds of pages. Most of them need magic that only our guidexml system can provide. So we need an evaluation if such a system fits our needs (which I doubt).

At any rate, I do thank you for your caring and proposals, but you need to consider the bigger picture here, many projects and people need to work together to make such a thing happen. Maybe start (yet another) discussion on gentoo-project.
I'm closing this bug as CANTFIX as it simply cannot fix the current situation.
Comment 5 Dan Douglas 2010-10-31 09:25:47 UTC
"very ugly" is subjective and unspecific. Readability and navigation are decent overall, though what was going through the head of whoever decided that neon green "Gentoo Linux" text on top of a purple background and logo with heavy JPG artifacting would be suitable for a distribution of the caliber of Gentoo is anyone's guess.

Are you thinking minor stylesheet tweaks, or major markup overhaul? Given that the the current layout is almost completely table-based, mixing and matching is probably not a very good idea.

There have been quite a few previous pushes to do some redesign. This old renovation project archived in the mailing lists seemed fairly mature and I'd like to borrow insights and good usable code from past work if possible. 

http://news.gmane.org/gmane.linux.gentoo.www-redesign/cutoff=136

Gentoo also has a lot of well defined backend infrastructure, document standards (like the XML stuff for the documentation project), and procedures to take into account. We'll have to listen to the rationale of whose who designed them and work around such things where necessary. Gentoo has pretty high accessibility requirements too so extensive CSS3 dependence probably isn't going to fly. People need to be able to navigate the handbook in Links etc.

That said, the code quality and look/feel of the website is significantly behind almost all other major distros. This reflects poorly on the distro as a whole; gives users and potential developers a bad first impression; and calls into question whether Gentoo is an actively maintained, modern, innovating operating system that they can depend upon for years to come. Some will argue that they like the "simplistic", "retro" look. Others will say that those who would judge a book by it's cover can go find themselves another distro. These are lame, harmful cop-outs.

Perhaps we can look at why previous attempts at improving the site have failed. Maybe if we try for smaller, more realistic incremental changes, and being willing to accept some transitional mess rather than reworking the entire infrastructure into a grand vision all at once would give better results in the long run.
Comment 6 Alec Warner archtester Gentoo Infrastructure gentoo-dev Security 2010-10-31 20:09:44 UTC
hi

(In reply to comment #5)
> "very ugly" is subjective and unspecific. Readability and navigation are decent
> overall, though what was going through the head of whoever decided that neon
> green "Gentoo Linux" text on top of a purple background and logo with heavy JPG
> artifacting would be suitable for a distribution of the caliber of Gentoo is
> anyone's guess.
> 
> Are you thinking minor stylesheet tweaks, or major markup overhaul? Given that
> the the current layout is almost completely table-based, mixing and matching is
> probably not a very good idea.

I have rewritten much of the site to not have tables anymore:

http://git.overlays.gentoo.org/gitweb/?p=proj/www-redesign.git;a=summary

I know nothing about web design and my XSLT sucks.  Plus I haven't worked on bits of the website in a while (last commit was six weeks ago.)  

> 
> There have been quite a few previous pushes to do some redesign. This old
> renovation project archived in the mailing lists seemed fairly mature and I'd
> like to borrow insights and good usable code from past work if possible. 
> 
> http://news.gmane.org/gmane.linux.gentoo.www-redesign/cutoff=136
> 
> Gentoo also has a lot of well defined backend infrastructure, document
> standards (like the XML stuff for the documentation project), and procedures to
> take into account. We'll have to listen to the rationale of whose who designed
> them and work around such things where necessary. Gentoo has pretty high
> accessibility requirements too so extensive CSS3 dependence probably isn't
> going to fly. People need to be able to navigate the handbook in Links etc.
> 
> That said, the code quality and look/feel of the website is significantly
> behind almost all other major distros. This reflects poorly on the distro as a
> whole; gives users and potential developers a bad first impression; and calls
> into question whether Gentoo is an actively maintained, modern, innovating
> operating system that they can depend upon for years to come. Some will argue
> that they like the "simplistic", "retro" look. Others will say that those who
> would judge a book by it's cover can go find themselves another distro. These
> are lame, harmful cop-outs.

I pulled the top 10 distributions from distrowatch.  8 out of ten have same modern layout (title bar, menu at the top, content centered with whitespace on either side.

> 
> Perhaps we can look at why previous attempts at improving the site have failed.
> Maybe if we try for smaller, more realistic incremental changes, and being
> willing to accept some transitional mess rather than reworking the entire
> infrastructure into a grand vision all at once would give better results in the
> long run.
> 

There are essentially three common problems:

1) The current system is XML + XSLT.  No one claims to know XSLT.  I'm happy to program in it but I am by no means an expert.  Luckily almost all the changes I desire do not require extensive xslt modifications.

2) The XSLT is complicated and no one knows how it fits together.  Making incremental changes is difficult because it is not clear what pages use what XSLT functions.  We do not have the testing infrastructure to test the changes on every page.

3) When folks propose new systems (Drupal, Django, Rails, whatever) there are always folks that bring up a few arguments:
  Load - We only have a few web nodes and our current content is fairly static (except for the adbar?) so we can do a high qps with just 2 nodes.
  Authentication - Shell access to CVS is required to add content.  No pesky security issues involving logging into webapps.  The ways by which dynamic content is added to the are simple and well understood.
  Security - The attack vectors for the current site are well understood; attack vectors for other apps (custom or open source) are unknown.  Any choice is likely to be less secure than our current infrastructure and cost more people time to keep up to date.

In the end I have talked to many people who are vaguely interested in a site re-design.  However none of them want to use the current infrastructure; however moving off the existing infrastructure is a giant pain.

Statistics on document hits would be useful in pruning pages we would need to migrate to a new system.  If a sizeable percentage of documents are not often touched we can move them to a legacy service (that continues to use GORG and XML documents) and migrate the trafficked pages to something new.
Comment 7 Dan Douglas 2010-11-02 02:55:20 UTC
Wow I had no idea about your project but it does look interesting. I know a bit of XSLT but will have to learn it much more in depth for classes next semester so I'll definitely be following this.

(In reply to comment #6)
> There are essentially three common problems:
> 
> 1) The current system is XML + XSLT.  No one claims to know XSLT.  I'm happy to
> program in it but I am by no means an expert.  Luckily almost all the changes I
> desire do not require extensive xslt modifications.
> 
> 2) The XSLT is complicated and no one knows how it fits together.  Making
> incremental changes is difficult because it is not clear what pages use what
> XSLT functions.  We do not have the testing infrastructure to test the changes
> on every page.
> 
> 3) When folks propose new systems (Drupal, Django, Rails, whatever) there are
> always folks that bring up a few arguments:
>   Load - We only have a few web nodes and our current content is fairly static
> (except for the adbar?) so we can do a high qps with just 2 nodes.
>   Authentication - Shell access to CVS is required to add content.  No pesky
> security issues involving logging into webapps.  The ways by which dynamic
> content is added to the are simple and well understood.
>   Security - The attack vectors for the current site are well understood;
> attack vectors for other apps (custom or open source) are unknown.  Any choice
> is likely to be less secure than our current infrastructure and cost more
> people time to keep up to date.
> 
> In the end I have talked to many people who are vaguely interested in a site
> re-design.  However none of them want to use the current infrastructure;
> however moving off the existing infrastructure is a giant pain.
> 
> Statistics on document hits would be useful in pruning pages we would need to
> migrate to a new system.  If a sizeable percentage of documents are not often
> touched we can move them to a legacy service (that continues to use GORG and
> XML documents) and migrate the trafficked pages to something new.
> 

The few people I've talked to also seem mostly interested in moving to a new framework like Django, but I think your way is appealing for the reasons you listed. Most of the content on Gentoo.org is textual and mostly static. XML/XSLT is really cool because people can continue publishing content in XML according to whatever schema they choose, and we can simply transform their data and present it to UAs however we choose (even in the case of scripts and dynamic stuff which spits out XML like the news and blogs on the front page, or packages.gentoo.org). No big CMS required, and the data is kept totally independent.

FYI We've set up an unofficialish Freenode channel ##gentoo-redesign. Not sure if there's already a better established place, but it might help to aggregate all parties potentially interested in working on website-related improvements. Feel free to idle etc. I'm sure there are competent web developers and XML hackers in Gentoo land who might be willing to lend a hand.