CompuServe Thread

Forum unknown · Telecommunications

#cocomm

15 messages in this thread
#133117From: Ben BlishJul 4, 1988 12:19 AM
No problem. The reason no one has done that is beacause the console device does not maintain the text is ascii mode (because the people who wrote it can't think their way out of a BCPL bag) and you'd have to figure out what is on screen by bit-bonging. There is a utility in the DLs that uses pattern recognition called "Snipit" that allows cut and paste…. hmmmm….. wonder if that guy could be convinced into doing a print screen? Anyone know who he is? I have long since thrown out the docs, if there were any with it. As for CoComm, sure. –Ben–
#133128From: Don Curtis/SYSOPJul 4, 1988 12:39 AM
Bit recognition works fine, as long as you use the default font. Once beyond that…you'd have to do the comparison between the bits on screen with the font file and then save the ASCII character. What's really needed is for the text routines to also write the ASCII character to non-displayed character mapped 'screen'. We all know how slow the text routines are right now…imagine how much they'd slow down if they also had to keep this 2nd memory area updatted including scrolling, clearing, cursor positioning and which window and/or screen was in front. In other words, a major can of worms.
#133158From: Henry ColonnaJul 4, 1988 11:13 AM
If you think the Amiga is slow at text – don't get near an ST. The Amiga is blinding compared to that pig.
#133216From: Don Curtis/SYSOPJul 4, 1988 5:26 PM
Henry, Oh…believe me I know about the slowness in the ST…working on the now defunct paperclip…we had all kinds of 'fun' trying to get it sped up. Dan did most of the work, but even using the line-a routines, it was still slow (no blitter). Things like early telecomm programs on the Amiga had problems with text speed since the authors used 'get a character from the buffer and print it to the screen' routines. They quickly learned that the proper way to go was 'get all the characters in the buffer and print them to the screen' routines. That sped things up significantly. Don
#133221From: Henry ColonnaJul 4, 1988 5:36 PM
A GEM window can't even handle 1200 baud without serious patching – but I guess you know that. 🙂
#133221From: Henry ColonnaJul 4, 1988 5:36 PM
A GEM window can't even handle 1200 baud without serious patching – but I guess you know that. 🙂
#133216From: Don Curtis/SYSOPJul 4, 1988 5:26 PM
Henry, Oh…believe me I know about the slowness in the ST…working on the now defunct paperclip…we had all kinds of 'fun' trying to get it sped up. Dan did most of the work, but even using the line-a routines, it was still slow (no blitter). Things like early telecomm programs on the Amiga had problems with text speed since the authors used 'get a character from the buffer and print it to the screen' routines. They quickly learned that the proper way to go was 'get all the characters in the buffer and print them to the screen' routines. That sped things up significantly. Don
#133158From: Henry ColonnaJul 4, 1988 11:13 AM
If you think the Amiga is slow at text – don't get near an ST. The Amiga is blinding compared to that pig.
#133294From: Ben BlishJul 5, 1988 1:15 AM
You, or anyone else, give me the source for a current, working console driver, and I'll make it work – and faster than the current driver, too, I bet. Handing 80×24 characters is _nothing_ compared to handling a complete bit map as well. For screen dumping, all you need are the character codes, anyway – which are being generated alla time. The console driver needs to work differently… not just have text handling pasted on top of it's current, badly thought out code. You need to manipuate the text – and then display it. That way you could have an (efficient) screen display that has some history in a realistic context – ascii, that is. Currently, no matter what font is in use, the font is in memory, right? ROM or RAM, what's the difference? Bit recognition seems like an _awfully_ poor way to go, (tho cute) for performance. I want a console that is 80×24, period. WAAAHHHH! –Ben–
#133321From: Vic WagnerJul 5, 1988 3:26 AM
What's so *&@*&#$^% magic about 80×24? I want 132×50!! Now what are you gonna do?
#133324From: Ben BlishJul 5, 1988 3:32 AM
Enlarge the array. <smug look> –Ben–
#133324From: Ben BlishJul 5, 1988 3:32 AM
Enlarge the array. <smug look> –Ben–
#133321From: Vic WagnerJul 5, 1988 3:26 AM
What's so *&@*&#$^% magic about 80×24? I want 132×50!! Now what are you gonna do?
#133294From: Ben BlishJul 5, 1988 1:15 AM
You, or anyone else, give me the source for a current, working console driver, and I'll make it work – and faster than the current driver, too, I bet. Handing 80×24 characters is _nothing_ compared to handling a complete bit map as well. For screen dumping, all you need are the character codes, anyway – which are being generated alla time. The console driver needs to work differently… not just have text handling pasted on top of it's current, badly thought out code. You need to manipuate the text – and then display it. That way you could have an (efficient) screen display that has some history in a realistic context – ascii, that is. Currently, no matter what font is in use, the font is in memory, right? ROM or RAM, what's the difference? Bit recognition seems like an _awfully_ poor way to go, (tho cute) for performance. I want a console that is 80×24, period. WAAAHHHH! –Ben–
#133128From: Don Curtis/SYSOPJul 4, 1988 12:39 AM
Bit recognition works fine, as long as you use the default font. Once beyond that…you'd have to do the comparison between the bits on screen with the font file and then save the ASCII character. What's really needed is for the text routines to also write the ASCII character to non-displayed character mapped 'screen'. We all know how slow the text routines are right now…imagine how much they'd slow down if they also had to keep this 2nd memory area updatted including scrolling, clearing, cursor positioning and which window and/or screen was in front. In other words, a major can of worms.