h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

I was working on the same idea.

You might want to strip out the tags, they're taking up about two thirds of the file and are the easieat way to reduce download time. You can also use larger blocks of colour , that would help. The more effective way is to combine widths and heights and calculate the largest block sizes but it's a pain in the bom to automate.

It's a shame that it won't work in Netscape unless there are single pixle solid GIFs on h2g2 that you can resize to form the coloured table elements. You'll need at least 2 to do a "black and white" image but you might be able to do something with re-sized smileys

Goof,

Thanks for the input... The tags are inserted on-the-fly by H2G2's parser; they're not in the original table. As for the wider blocks, I have a half-working version which uses "Colspan=x" to achieve the same effect. If I could be bothered, I could look out a "best fit rectangles" algorithm I wrote back in my Amstrad CPC-464 days which recursively found the minimum number of rectangles required to recreate an image, and convert that to use and ...

The original Amstrad program was an experiment in image compression, iteratively drawing "rectangle approximations" of an image using different levels of detail, XORing them with the original image and performing run-length compression on the results. The object was to find the optimum balance to allow me to store a "rectangle list" along with the differences from the original image. I gave up on the project because a single small image was taking between three and five days to process...

I might ressurect the idea someday now that machines are a *little* faster (), using "Dark BASIC" or somesuch, and using LZH compression instead of run-length for storing the differences; hey! - I might just append a GIF of the differences to my rectangle list, and XOR it in after the rectangles are drawn!

I thought that might be the case with the font tags. h2g2 should really be produsing cleaner html

I wrote a program to produce those (now famous) 3D pictures that you can see without special glasses in 6502 assembler on the old BBC model b. That was a bit slow too
but nowhere near that bad.

I'm looking into some other ideas but am not really up to speen with what I can get away with GuideML-wise. I'll check back here now and then to see how you're doing (better than me no doubt)

Actually, Goof, I've been reading the Italics' replies to the propoganda, and some of the restrictions are, while still irritating, starting to make a little sense.

The BBC plans to make this service available on a number of platforms, with different limitations in resolution. I suspect that's why they don't allow an absolute size to be specified for the "Blob" graphics; they will have a different set of source files for each target device, at different resolutions and colour depths, and all pages will need to work with all versions. This is also the problem with external graphics; unless they are viewable on all platforms, they reduce the value of the site to the BBC as a "technology showcase". They're less worried about linking to copyright material than they are about the fact that they would have to make copies of such material at different resolutions for, say WAP, WAP-2 or Web-TV... This has been pieced together by "reading between the lines" of various fora, and is a personal view which is not the view of the BBC, unless specifically stated. I challenge them to specifically state something!

What we need is for them to embrace Flash vectors, and let us upload our own scalable graphics...

Hey, Goof... Check the page now; I suspect it won't be around for long...

Didn't understand a word of those explanations / comments, but the result is veeeerrrry impressive.
Congrats.

I thank you... I have to exercise my muse on occasion...

Howdy, Peet. So the upshot of this, if I understand correctly, is that the GuideML is set to automatically show tags as broken, but this background tag thing will still work at showing offsite images? Interesting.

Pretty much. But please don't do anything "silly" with it... Yet...

To use this knowledge effectively in a protest would require letting everybody know about it. I assume the moderators will be able to remove pages with images posted this way anyhow. And I really want to hold off on major protests until after we see whether the petition gets any results. If we get no response, or if we get the same kind of response as before, then will be the time for protesting, boycotting, whatever might work.