I've been thinking about this a lot recently. A lot of people redesign their sites all the time. Nobody is satisfied. What makes a really good website, from a geek-design point of view?
I wrote a while back about skinning websites. I want to expand on that, and go deeper, and get a lot more prescriptive.
There are possibly three areas that I want to talk about:
I have already discussed the skinning of a site, although my thoughts have now gone a little deeper. The various methods available all have repercussions on other aspects of a web-site, and these need to be addressed. I am beginning to tend ever so slightly towards Tristan's point of view that implementation details (even to the point of "is this page a directory or not") should not have any visible effect whatsoever (largely influenced, I think, by his use of Nevow, which stores pages as python functions mostly), but nowhere near that extreme. I'm not sure that pages should map one-to-one onto the filesystem, although there is definitely ease-of-maintenance potential in the filesystem that should not be discarded. I will write more on this later, when I have examined all the options.
Standards compliancy is something every geek knows about, but I am talking about 100% runs-to-the-core standards compliancy. I'm talking about using Unicode, and sticking to accessibility guidelines, and having meta-data, and correct CSS. But, more relevant to my "how to skin a site" topic, I want to make this standards-compliancy default in a web-site. With a few tweaks here and there, all the meta-data and so on will be put in automatically for every new page. Sure, you'll still need to know that an anchor-tag can't go inside another anchor-tag and so on, but a lot can be done for you.
The final point overlaps somewhat with the second one. A lot of it is covered in this checklist for web standards, but there's more to it even than that. An indication of what I am talking about can be gleaned from this rant on typesetting. I want quotation marks to be used correctly (even if it means using CSS "contentbefore" and "contentafter" properties to adjust the <q> tag), and so on. I haven't delved deep into this topic, but I'd like to look further. Russell knows a lot more on this topic (and most of the other issues I've mentioned in this page), but I'm going to do some reading and see.
I still need some sort of logical flow for this issue, some way to tie it all together. I'm hopefully going to start an outline soon, and then fill it in slowly as my thoughts come together.
I am back from SATNAC, my first conference, and it was interesting. It was fun, and it was revelatory.
The conference has been well written up by Ingrid, Darb, and Dom, who were blogging it as it happened for the most part. I am going to give a sort of overview and opinion here.
As far as the content of the conference goes, it can probably be summarised as follows:
"Old" PSTN (i.e. telephone) networks are on the way out. They will be replaced by the "Next Generation Network" (NGN) (remember that catchphrase, use it whereever you can), which will be an IP network, with MPLS (Multiprotocol Label Switching) running on top of it. This means an Internet network with lots of different protocols running on it. The old PSTN networks will, of course, not just disappear, they will be merged into the NGN slowly. We call this "convergence". The NGN will have lots of services available to its subscribers, using all the cool media like voice, data, video, and so on. Mobility will become rife, and it won't matter where you are or what device you are using, you will be able to access everything anyway. Wireless computing will be everywhere, and all the operators are going to move into this new era with their heads held high.
And that's pretty much it. That's what everybody was about. Sure, there were papers about fuel cells, and security, and so on, but nobody really cared about them. Jason's VoIP PBX paper was packed to the brim, people sitting on the floor - everybody left as soon as it finished, and a paper about "geospatial data" was scheduled. There were catchphrases and acronyms floating about everywhere. Most companies used the conference to say things like: "This is where technology is going. It's awesome. And, by the way, our product is perfect for where this, and you should buy it".
There was quite a bit of Open Source hype, which pleased us. Nothing new, but it was clear that people were beginning to get it. Some of the questions at the open source talks were irritating, because they showed that a lot of the fallacies surrounding open source (it's not commercially viable, etc) were still around, but in general, there was a clear indication that people understood that supporting open source is supporting South Africa.
Other highlights of the conference were the enormous big cheese super-suit CEO of Telkom telling our own Professor Clayton that Rhodes was a crappy third-rate institution (his exact words were, in a debate chaired by Prof Clayton, "I was pleased to see the quality of work, not only from the premium research institutions like Wits and UCT, but also from places like Kwazulu-Natal, Fort Hare and Rhodes"); my own wonderful gaffe in mentioning to another suit from Telkom that we'd been using VoIP (contrary to South African law) for a while; and the fairly attractive young female open source advocate from the CSIR being mobbed by a bunch of drooling geeks after her talk.
On the social side of the conference, I didn't do as much "networking" as Darb did - he really got into the "meet the players in the field" thing. I spoke to a few people, met some geeks from University of Pretoria, but that was it.
The food and the "social events" were fantastic. Even if the rest of the conference had been kak, as Lorenzo put it, "at least we got food for a week and came to Stellenbosch".
The first night had a big buffet meal in Mojo at Spier, with awesome food, from ostrich to mussels to calamari. Claire came down with Random-Friend-Thandi, and we went out for a drink. We got lost on the way there, and had to ask a policeman to drive us to Academia House (to drop Mamello off first). Then, on our way home, we got lost again. We drove around Stellenbosch for an hour and a half, heading out one way, and out another, and just could not find where we wanted to go. In the end, we stopped another police car to ask for help and it was the same bloody cop. He must have thought we were nuts. We told him that we were from Joburg, and he seemed understanding.
The second night was a meal paid for by the CoE that cost more than four month's rent, but was enjoyable. Certain members of the party got hideously drunk, but it was fun.
The third night was the big farewell party at the Stellenbosch Town Hall. It was "district six street party" themed, with all sorts of Malay food, and a lot of drunken aged nerds trying to be cool. They square-danced. We smuggled wine and brandy out, very surreptitiously and sneakily. The bouncer told us later that he didn't actually mind. Somewhat deflated us. Claire and the marvellous Adeline came down and joined us, and we went out for more drinks. I didn't really get to spend that much time with them, but it was fantastic to see them again. I may take Ads's advice and become a rent-boy in Cape Town, so I can see them more often.
So, we are back in Grahamstown. I left my souvenier mug in the bed-and-breakfast, but I managed to bring the pen and the bag back, so the conference wasn't a total waste.