[Coco] Coco 3 and the Internet

[Roger Taylor]

At 08:11 AM 6/5/2004, you wrote:
>At the very least, it should be trivial to design a simple HTML parser 
>which would get HTML text from some input path and call "tag handlers" to 
>handle the various tags in the text.  This can be done now, simply by 
>using test HTML pages on a floppy or hard drive.
>
>It would be tedious to write in assembly, but would probably offer the 
>best in terms of speed and code size.  I say someone should take up the 
>challenge and start on one!
>
>Boisy


Sock's Hi-Color trick would be great for this.  That, or my old/unreleased 
Net-Mate video that does 8 colors per foreground/background on the graphics 
screen.  It prints all of the "IBM" graphics characters/lines/angles, etc. 
but the text font is kinda compressed.

It would be very easy to render simple HTML content, but most stuff on the 
web today would choke a weak browser.  You'd just have to ignore tags you 
couldn't process and hope for the best.  Images would be impractical to 
support, especially since we've never pull off the JPEG format, and GIF 
does take a while to render.  Animated GIFs would be hard to display.

Another thing to consider is the height limit of a page... some pages just 
keep going on forever it seems.  To support this on the CoCo, you'd have to 
render directly to a virtual screen that runs from block #0 to #55 or so, 
or perhaps block #64 to #255 on a 2-meg machine.  Blocks #64 to #127 are 
free (in Disk BASIC) on a 1-meg machine.  You'd scroll through that memory 
just like the PC does when you scroll up or down on a tall web page.