Slash Boxes
NOTE: use Perl; is on undef hiatus. You can read content, but you can't post it. More info will be forthcoming forthcomingly.

All the Perl that's Practical to Extract and Report

use Perl Log In

Log In

[ Create a new account ]

Alias (5735)

  (email not shown publicly)

Journal of Alias (5735)

Wednesday April 22, 2009
12:28 AM

Beautiful is better than ugly

[ #38842 ]

The Zen of Python is a set of 20 aphorisms that "succinctly channels the BDFL's guiding principles for Python's design".

Some of these are contradictory, or consist of pairs saying "This is true. But not really". But we can ignore all that.

Because one of them stands out clearly at the top.

Beautiful is better than ugly

Beauty is something Perl does really badly, because the Perl community as a group trends towards pragmatists and sysadmins. We lost the "front end" (HTML-side) of the community to PHP long ago, and never really recovered.

Perl has a public reputation of having ugly code. This can be true, but is often incorrect or besides the point. But with a stereotype like this in play the LAST thing we want to do is have ugly websites, which only serve to reinforce "Perl (code) is ugly" to the uninformed.

This is especially true in the case of our oldest core websites. is horrible.

An 800-wide fixed-width site with dark, block colours, small type in a link colour not clearly distinguishable from black to non-perfect eyes and purple "visited links" ala 1998 and dense clusters of links with no descriptions, ratings or ranking.

It's slogan is "When you need perl, think".

But clearly not many people do, with the number of distinct visitors to alone running 1-2 orders of magnitude higher than

The "download @ % Perl" image link is ugly and doesn't even look clickable and the yellow box that says "Where do you want to go tomorrow?" is both meaningless and out of place.

Even as a "directory", the website fails horribly to help me with anything I want to do. is almost as bad.

It's minimalist, contains almost no information, and itself only serves as a directory to other CPAN websites.

But of course it also has a website design circa 1998, doesn't actually link to most of the CPAN-related websites at all, and doesn't even have a search box! You have to read through three search website descriptions and pick one before you can actually search.

And lets not forget, which has (admittedly very slowly) degraded over the years under O'Reilly's stewardship.

The irony in this situation is that the most useful link on is the link to "our expanded news coverage" at, which is a far far better website than Even though everything on is based around commercial services offered by O'Reilly, it is a much more useful resource for commercial Perl people when compared to

When someone turns up to the, or websites with some task in mind, we don't do much of anything to help them get where they are going.

Looking through the list of all the other Perl websites, the look and feel disaster just keeps going on and on. is just a three page ugly clone of the ugly website, with two of those three pages links back to is another three page ugly mid-nineties style website, although this one isn't a copy of It's a whole different ugly. is simple, clean, suitably elegent and has a search box right on the front page, where you'd expect it. And, I'm sure partly because of this, it also has the highest traffic of any Perl website. is another ugly copy of the website, but feels more comfortable in that skin, because it's behaviour more like a simple review blog instead of an actual website, and it points you in plain language directly at the three things you might actually want to do when you arrive there. isn't too bad, but has a completely different colour scheme from the rest of the Perl websites (most of which are blue) and is to some extent "confusingly similar" to

Surely it would be nicer if some of the annotation Javascript magic was just merged into search.cpan, and you only went to the main annocpan site when you REALLY need to dive into annotations vs revisioning. is another ugly clone, with a front page that's basically just a joke (although hey, it is at least informative). is different from BOTH the style and the style. is horrible, but at least it knows that it is and is going to change. is reasonably competent (although you'd expect that from a group of people who are completely web-focused).

The YAPC website at is decent, which you'd hope for a conference website who's entirely reason for existing at all is as marketing. O'Reilly's OSCON website is of course great (although we've finally reached the end-state, where the string "The Perl Conference" does not actually appear anywhere on the website, that Google can see anyway). is another horribly plain website with a front page image circa 2000, although at least it's not fixed width. is actually pretty good looking, and does look like something you'd expect from a document archive. In that sense, it has a similar suitability-for-purpose feel to the Oracle documentation archives.

Except, of course, doesn't actually have all the Perl documentation like the typical visitor would expect.

At the Birmingham hackathon Jon mentioned something about this being because he builds the website on his laptop, and it would take too long to upload via FTP/rsync if he built it for all versions of Perl.

Can someone please hunt Jon down and force him to accept your donated server space to generate the full website on? As one of the few people we have that can produce a good looking and functional website, we really should be giving him more technical/infrastructure backup. is fucking horrible and looks like the kind of Windows 3.1 post-install help pages that were being generated in 1993. is sparse, but functional, although it could do with someone giving it a little chrome.

And of course, nobody has any idea that it exists. Partly because we also have which is a second website for the same mailing lists. No, I don't understand the difference between the two either. is fine, but someone needs to have a talk to Brian about some of those yellow-on-white mouseovers. is mostly ok, but as it mentions on the front page has been podfading a bit, and doesn't release on a regular schedule that people can anticipate. is crufty and could use some CSS attention. But it is a great community and does seem to have made something positive out of the cruftyness. After all, monks themselves are pretty crufty. Of course, I still find the chatterbox weird and hard to use, and if I can't use it as an semi-regular visitor how is any casual arrival going to use it.

Apart from the ugly chrome actually does a really good job of aggregation. I have a small question about why Plagger wasn't good enough to run our Planet website, but that's a secondary issue and something I guess I'll take up with Miyagawa one day.

And of course there's the debacle. It's now mostly a news site without much news, hosting a highly valuable set of blogs that can't do images, syndicate to the blog aggregation services, support trackbacks or do any other normal bloggy things that help get our message out into the wider community.

Given the astonishingly large amount of blog software that is written in Perl, I find it astonishing that no Perl blogging companies, and no Perl blog system authors have managed to find the time to make a best-of-breed

I suspect that if we had such a thing, most of the people still on (including me) would move over to it pretty much overnight, and then can go back to what appears to be it's real purpose, being pudge's private beta testing website for slashdot. is also another ugly website, and doesn't contain anything on it that actually tries to convince me to give anyone any money.

The Perl Foundation's website is just as horrible, although there's another bitter irony here that when the Perl Foundation News section basically failed, they created a "blog" (called which is really a news page, and it's far far prettier and usable than the actual Perl Foundation website (although it's still fairly plain and ordinary by normal standards).

If we look at the CPAN websites, things are just as bad. has been improving in leaps and bounds functionally but the look is horrible... except for its wiki at which looks pretty good, but is really mostly for internal consumption by a small number of people. And the statistics website is yet another ugly clone.

The CPANTS website is also horrible. The name CPAN Testing Service remains annoyingly and confusingly similar to CPAN Testers and badly needs rebranding. Perhaps "CPAN QA"? (but keep "Kwalitee for the name of the metric).

And did you even know that the CPAN had a bookmark database? is another practically useless directory, but this one is so bad we could live with it just dying outright. It's link to doesn't even point to a valid page. is so simple, trivial and ugly that even my personal repository at, which I INTENTIONALLY made ultra small and ugly so it would work on my crippled mobile phone browser, looks good by comparison. What does it say when someone trying to look ugly can end up far better looking and functional than the official Perl subversion repository.

And hey, at least my list of modules doesn't throw Python exceptions like the one does.

I could keep going (I probably should have stopped long ago) and I'm not even going to bother talking about technical beauty yet in this post.

What's it going to take to fix this crap?

Do we need a dedicated website group who can start taking over the crap and come up with a standard website design that doesn't look a decade out of date?

Do I need to start taking over these websites and moving them into my repository? I already have a collection of six simple websites that anyone with commit to my repository can edit, and they are automatically synchronised from SVN to the web server. Should I just start sucking in all the small websites?

Could I just take over and all those bits and pieces websites and act as a steward in the same manner I've done for abandoned toolchain modules and (most recently) DBD::SQLite?

Would a Vertical Metre of Beer competition for a new look and feel for all of the Perl/CPAN websites be useful?

And for god sake, can someone with a spare thousand bucks leftover from a YAPC PLEASE organise legitimate crypto certificates for the SSL websites? This whole self-signed certificate crap makes us look like incompetent amateurs (when we are supposed to actually be competent amateurs)

The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
More | Login | Reply
Loading... please wait.
  • ...would be to get a professional designer to create a real logo and a site template, imo this is long overdue.
    And by professional designer i mean someone prominent, like Jon Hicks.
    • The logo we already have. The onion can be used really well and in different ways.

      It's the site template, done in a way that we people making up new websites can easily integrate, that we badly need.

        • Most of Andy's changes were meaningless cosmetic changes, and I simply don't care about that, and don't have the time to give to shepherding through things that to me are unimportant. I have too many actually important things to ignore! I know a lot of you care deeply about how shiny a web page is. I am not one of these people. I don't apologize for it, and I find it odd that your perception of the site would be so drastically impacted by that.

          As to the lack of news that Alias mentioned, I've many times

          • Oh and BTW, I think Andy's changes mostly look good, I am not attacking him or them; I am just saying it's something I don't have the time to care about.

              • Perhaps. As noted, right now the code on this site is a bit behind. I am planning on updating it soon, within the next few weeks. I would be open to pure CSS+images changes. Unlikely I'll be too interested in code/HTML changes, as those tend to take up a lot of time, but you never know.

            • You appear to be trying to criticize me, but I don't think you actually are.

                • Nope, you've confirmed I was right all along. You were not, in fact, criticizing me. You were trying to, but you did not. You think it is a criticism of me that I don't care about such things. It's not. You think that the fact that I don't care about such things means I "have my head up my ass." You are wrong.

                  I could, of course, say the exact same thing to you, from my perspective: anyone who thinks a "shiny" web page makes a lick of difference, and actually cares about such things, has his head up hi

  • I whole heartedly agree that the websites for the CPAN Testers family has been in need of a complete overhaul for sometime. I have asked occasionally of designers I know for some ideas, but so far nothing has stood out. The Wiki was my rework of an OSWD [] template, which originally was intended to be the first branding of the CPAN Testers websites, with the others using the same layout, but different colours. However, I didn't think it suited the other sites.

    The current CPAN Testers Reports site makes use of

    • Actually having a quick look this morning, there is a recently uploaded design to OSWD that fits very well. Clean, professional and functional. I'll have a go at reworking for CPAN Testers :)

  • Given the astonishingly large amount of blog software that is written in Perl, I find it astonishing that no Perl blogging companies, and no Perl blog system authors have managed to find the time to make a best-of-breed

    I was astonished by that too. But having prodded round the edges of the problem, it seems to me that no Perl blogging engine is up to the job.

    If anyone reading this thinks I'm talking nonsense and would like to prove me wrong, then please get in touch.

      • but what about Movable Type?

        That's exactly what I thought. I tried. It didn't work.

        I'm hoping that this discussion will flush out an MT expert who will be able to show me where I went wrong.

          • Oh it installs just fine. I'm really happy with MT. I run many blogs on it.

            But this was the first time that I had tried to set up a multi-blog installation. I assumed that it would all just work out of the box, but it didn't.

            There are a couple of discussions that I had on the MT forum here [] and here [].

            There was a new release a couple of weeks ago, so I might see if I can find the time to try again over the next week or so.

            Oh, and by the way, I run all of my MT installations on 1&1 servers. Not sure why you

              • I don't think you need shell access to set up MT, but I've always done it using shell access so I could be wrong there.

                I have a root server with them. I can't imagine running web sites on a server that I didn't have full control over :-)

                  • MT bundles all of the CPAN modules that it needs. So if you can FTP a directory tree into your web space then you're all set. All of the set-up and configuration is now run through a (very nice) web-based system.

                    I have a Value Server [] from 1&1. It costs me GBP 30/month and I happily run about a dozen [] domains on it.

      • it needs buy-in from the appropriate people

        Which also means some technical constraints*, because it has to be able to run on the existing hardware budget if not the existing hardware, and scale to the size of known traffic peaks. Yes, these sites can end up on the front page of slashdot, so the "static" pages better not be making umpteen calls to a database to generate their content.

        * constraints are good. A good artist uses constraints to channel their creativity and inspire them. Well, that's my story an

          • I think you're missing the point. Adam is talking about presentation and layout design, not a CMS or whether we're using a flavour of the month framework. Templates and CSS can be integrated into any site. Ruby on Rails largely got peoples attention because the initial websites looked well designed, and had nothing to do with the backend codebase.

            The problem is that many of us are decent coders, and can put together a functional site pretty well. However, we're mostly not website designers and that's what w

                    • I have intended The Perl Beginners' Site [] to be the (unofficial so far) first-stop for Perl beginners. About two years ago, I've adapted an OSWD template for its look, which should have made it attractive enough. There's still some problem with the testimonials section being lacking, but I plan to correct it when I find some time to dedicate to it.

                      Contributing and building upon Perl-Begin is very easy because the sources are available in a publicly-accessible Subversion repository, and the licence is CC-b

        • Beauty is better than ugly...sometimes.

          Perl is good, the marketing side is ugly as Alias has pointed out.

          To be honest, the glass onion look is a great start in my opinion.

                • Some people say "Perl is dead". If I didn't know where to look, the Web sites I find would imply "Perl is dead". Appearance matters. I think that's what Adam's trying to say. I agree with him 100%.

                  No one is saying we should be Python. No one is saying that we got into this mess intentionally. No one is saying anything other than "our Web sites look amateurish and this has an impact on perception". We do have a perception problem and we need to address it.

    • Yeah, the comment system is astonishingly bad. When I first visited this Web site when I wasn't logged in, I was shocked at how unusable it was. I even switched browsers on the off chance that it was a browser bug. It's not :(

      • Not for nothing, but the comment system works astonishingly well for Slashdot.

    • Is there no other traffic because no one is interested in anything else or because people don't find the site useful?