From the Orient

I’ve discovered that Apache doesn’t generated the headers Last-Modified and Etags for server-parsed documents [this is probably old news for everyone else]. This means they are never cached. You can work round this by setting XBitHack to full in your .htaccess file as discussed by Bowie Snyder. This means you can generated a Last-Modified header, but I’ve found no way of generating ETags for SSI documents.

This is quite important for a couple of reasons. Firstly, as I’ve already mentioned, the documents cannot be cached meaning more download times and greater bandwidth.

Secondly, RSS aggregators are starting to use these headers to control RSS bandwidth issues. If XHTML syndication ever happens (I see it happening as a choice alongside RSS), it is probably not advisable to use SSI.

As I only use SSI to build pages out of other bits of pages, rather than provide any dynamic content, it is really not a good idea for me to be using SSI, because I can do that from my MT templates. I think it is now time for me to learn mod_rewrite.

BDG to Etags

Paul Internet

I’ve been working on a little idea in the past couple of days to try and managed my multiple blogs a little better. Whilst every blog has its own page, every attempt to see everything on one page falls short in some way.

This is the latest attempt. Using CSS and a styleswitcher to control what is displayed. It uses the styleswitcher from A List Apart and therefore required no coding on my part. Multiple blogs are included on one page using SSI and there are multiple style sheets to determine where you see what.

It works quite nicely once loaded, and it quick to change and it remembers what you last chose using cookies. For people without javascript, the links take them to the relevant page. [The HTML template is virtually that used by diveintomark. It thought I'd take that as a default and modify as necessary, but it seemed too good to modify. And this is only a proof of concept idea anyway.]

I’v decided though that this is a flawed idea

• The page takes too long to load and contains more information than is displayed, so wastes bandwidth
• Requires javascript, and if you are going to use javascript, you might as well use javascript and DOM to control this ala ReadingEd [Not that I can program in javascript and DOM]
• Making a simple CSS change becomes more complicated with there are multiple CSS files to modify.
• Without a very large number of CSS files, you are limited to what you can do. You can only have a certain combination, or a huge amount of links to maintain.
• You could get round the point above by programmatically writing your CSS file on the fly, but that too is pointless because if you are going to do that you might as well use DOM again.

If anyone has any comments (good or bad), please let me know.

Paul Web Coding and Design

I didn’t quite last a month, but I decided to start putting entries into this blog again. I don’t think I’ve changed my focus at all in the end. I should point out I’m strictly amateur in most things I talk about. But the best way you can learn is by investigation and playing with the technologies. And if I have said something wrong, and I’m told so, that is actually a good thing because I’ve learnt something. And maybe, just maybe, they are wrong themselves.

So if you see my opinion as wrong, or just plain dumb, for goodness sake, let me know. I won’t be offended … again.

Paul Blogging

Mark Pilgrim thinks the world will win in the case of XHTML vs. the world and has written a python program to turn tantek’s xhtml into RSS. I’m thinking this has proved Tantek’s point; XHTML is versatile enough to be syndicated, so why bother producing an RSS feed? Simply add XHTML parsing into today’s RSS tools. And the next step will be to cut RSS (thanks for spotting that Albert) out of the loop and have tools that parse XHTML directly.

The rebellion started with Anil Dash suggesting XHTML as a new syndication format, and was taken up by a variety of other people. A summary of what can be acheived now has been collated by Aaron Swartz, including Site Summaries in XHTML.

Ian Hickson descibes it nicely

You see, people are finally catching on than XHTML has actual semantics (sound familiar?), and thus software could be doing a lot of the content aggregation work for us.

Some questions remained.

‘Do you lose something if you use XHTML over RSS? The abbreviation of RSS content for example?’ Tantek: ‘You can minimize your top level XHTML content just as you can abbreviate RSS content, and have “More…” hyperlinks on all your entries if you wish.’

‘Don’t you reduce server load by using RSS?’ I can’t see this changing because the same people will be using XHTML syndication to hit the server as use RSS. What is the difference

‘RSS is for machine processing while XHTML is designed for display.’ Scott Andrew (in the comments of Anil’s post): ‘Properly-structured XHTML is far more robust than RSS for providing syntactic structure for a Web document, and is just as machine readable.’

Further reading

• scottandrew.com
• as days pass by : XHTML instead of RSS
• tantek/log : More on XHTML syndication
• as days pass by : More on XHTML vs. RSS
• techno weenie : XHTML syndication

Paul XML

The Ghost of Computing Past

Paul has a dialup connection to the Internet using a PC, no mobile device with internet connection, and there is little or no information available on the internet regarding Paul’s journey to work.

Paul oversleeps, gets up, and jumps into his car to begin his journey to work. He hasn’t bothered to look on the internet at the travel cams, as it just takes too long too start the computer, dial up and get the information. Just as he gets on the motorway, a traffic radio report comes on and informs him there is a major accident on the motorway causing severe delays. Two and a half hours later, Paul arrives at work, frustrated and tired before the day begins (Total journey time; 2 hours 45 minutes).

In the evening, Paul blogs about his frustrating journey into work.

The Ghost of Computing Present

Paul has an ADSL connection to the internet using a PC, has a Sharp Zaurus, but no wireless card so it has no internet access. There are some good traffic sites on the internet, but no standard format or system exists.

Paul oversleeps, gets up, checks the BBCi website and discovers chaos on the roads and rail. A major accident on the motorway is causing all sorts of problems and a derailed train is causing rail problems to and from Paddington. A look at an online train timetable shows that the trains are only running five minutes late and the problem at Paddington doesn’t look like it is going to affect him.

The train arrives at the station at the time the website indicated, even though Reading Station is in chaos due to the problems at Paddington. Many people there are asking Railtrack officials who seem to know less about the problem than Paul.

A slightly smug Paul exits the station, but he is slightly crest fallen to see an unusually large crowd of people waiting at the bus stop where he needs to get his bus to work. However, about five minutes later the bus arrives and he feels happy. However, the traffic in the center of Reading is appallingly bad and it takes nearly an hour to get to work, a trip that normally takes ten minutes. Paul has to stand the entire journey. Paul later discovers a fire closed a road and caused further travel problems. (Total journey time; 1 hour 45 minutes)

In the evening, Paul takes the opportunity to type up a report of the days events using his Zaurus on the bus trip home. When he gets in, he copies the file over and creates a blog enty with it.

The Ghost of Computing Future

Paul has an ADSL connection to the internet with a wireless network installed. He has a Sharp Zaurus with a wireless card attached and has written some program in Python to access a variety of travel web services.

Paul oversleeps, gets up and notices a flashing light from his Zaurus sitting beside his bed. It is a couple of alerts from a his Travel Service program. An accident on the motorway and a derailment at West Ealing Station. Paul checks the latest train time from his bed and discovers the next train is only running five minutes late and decided to go by rail. He notices that he has enough time for a coffee before leaving to catch it.

After a successful journey to Reading, he exits the station only to find an unusually large number of people waiting for the bus. He checks his Zaurus and discovers that whilst he was on the station platform, his Zaurus has connected to a wireless LAN and grabbed some more travel information and there is an alert from his program. His Zaurus tells him of a fire in central Reading. He decides to walk; a pleasant half hour stroll alongside the canal. (Total journey time; 1 hour)

Paul types this up on his Zaurus on the bus on the way home and when he walks through the door, it is automatically posted to his blog.

Paul Internet

From the Orient is on hiatus. Meanwhile, enjoy the list’o'links, or head to Vent of Hearing for something else.

Paul Blogging