K538915@CZHRZU1A.BITNET (03/18/88)
In article <496@csvax.liv.ac.uk> sqrkl@csvax.liv.ac.uk (Richard k. Lloyd) writes >Firstly, Simon, I'm sorry about posting this up to the net, but I'm >not convinced that I have access to your UUCP/Bitnet addresses... You do, just try.... ........... >>If you want to design 6 fonts, three in a 4x4 cell, and the other half >>in a 4*2.5 (:-)) cell, your welcome. Seriously, designing fonts is a >>LOT of work and adding more fonts means the program gets larger >Obviously, this is a limitation of the Atari ST. I wrote a font designer >program for my Archimedes and designed a complete (256 char, including >the digital chars) character set that uses 8 * 10 pixels in about 2 hours. >I don't need 6 fonts because I've written ARM routines that support all >double height/double width modes on a VT100 (that took about 2 days of >coding). I think that 6 fonts is over the top - you should be writing >code to expand the original set (so you need only 1 font). You're jumping to conclusions again....the scheme I use in UniTerm is: HiRes: Font(w*h)| 8*8 | 8*16 | 16*16 | 16*32 | 5*8 | 5*16 | 10*16 | 10*32 | ---------|-------------------------------------------------------------| Normal | N8*8 | N8*16 | N8*8 | N8*16 | N5*8 | N5*16 | N5*8 | N5*16 | Special | S8*8 | S8*16 | S8*8 | S8*16 | S5*8 | S5*16 | S5*8 | S5*16 | Intern. | I8*8 | I8*16 | I8*8 | I8*16 | I5*8 | I5*16 | I5*8 | I5*16 | The 8 pixel high fonts are used for the 49 line mode, the double height/ width fonts are generated by from the smaller fonts, each of the character sets are a full 128 characters. Transforming the 5 pixel wide fonts into 8 pixel wide fonts justs looks too ugly. MedRes: Font(w*h)| 8*8 | 16*8 | 16*16 | 5*8 | 10*8 | 10*16 | ---------|-------------------------------------------| Normal | N8*8 | ? | N8*8 | N5*8 | ? | N5*8 | Special | S8*8 | ? | S8*8 | S5*8 | ? | S5*8 | Intern. | I8*8 | ? | I8*8 | I5*8 | ? | I5*8 | Since the trick with magnifying the font can't be done here, 3 16*8 fonts and 3 16*5 fonts would have to made, something that I haven't done mainly because of the additional ~10 kB memory (and size of the executable) this would use. [Just a remark on the side for the benefit of anybody still reading this, a number of people have asked me why their 'font-changers' don't effect the font UniTerm uses as 'Normal' font, UniTerm gets the address of the two system fonts it uses via the line-a variables, if the 'font-changer' changes these values everything should be ok. If the 'font-changer' changes the fonts the BIOS uses, you're out of luck, since UniTerm doesn't use the BIOS for character output ] >>The vector >>drawing is fast enough for 9600 baud, at least I've never had any >I disagree. TERMINAL/ST version 2.1 has vector drawing at about twice the >speed of yours (it was written in this department by a lecturer). Richard, when you measure something, you say how you mesured it and what you mesured. As an example: are you comparing drawing speeds using the playback function or the zoom mode or just simply with vectors being sent by a host computer? Anyway there was an error in the code that made the vector drawing in playback mode and while receiving from the RS232 port slower than necessary (but still > 2000 byte/s, that's why I didn't notice it), the new cursor postition was calculated after every received byte, instead of after each complete vector. (BTW I didn't claim it can't be done faster, just that it's not so slow that it is something that would have to be improved tomorrow) >>If you hold down the <Alternate> key, you've essentially got a one key >>NO SCROLL key, anyway in the more recent versions you can actually put >That's funny. I'm running UniTerm V2.0a 020 and the Alternate key doesn't >do an awful lot (I haven't reconfigured the keyboard in any way). You naturally still have to toggle it on/off with <Z>..... >DCL on a VAX/VMS system already has an SLE, so I disabled it...with annoying >results if I accidentally hit INSERT (this happens on V2.0a 020 as well). ...........and I didn't say that it doesn't happen with V2.0a 020, did I? >>Wrong! No VT1XX, VT2XX or VT3XX compatible terminal works this way, >>all clear the screen and home the cursor. The command set of D*C >Again, rather strange. I am using a PT100 (Plessey VT100 compatible) and >I can quite easily switch in and out of 132 column mode without a clear >screen/home cursor (although the terminal can be configured to do so via >the SETUP options). With other words it isn't VT100 compatible, is it? >Where does this mixed length business fit in ? I've never seen a terminal >that has different screen widths on the same display at the same time ! May I quote your original letter: |6) Retain the screen display when switching from or to 132 column mode, rather | than clearing the screen and homing the cursor. and now the Oxford Dictionary: retain .....keep in place, hold fixed..... Sorry, but you didn't say 'adjust' or 'truncate', you said 'retain', and if you 'retain' a 132 column line it stays a 132 column line. As I've said before truncating or adjusting the display is trivial, even if not standard and I might implement it if there's enough demand. >Again, V2.0a 020 DOES NOT switch the mouse on when it is moved. The I-bar >only appears when you click on the mouse button first. ...... I never said it did, did I? Richard, it might be worth getting used to the fact that you don't always have the latest version of a program and that people don't just fall over dead because the grand majestro wants something. >I've got the 'intelligent listen-back' cursor movement routines in the >Macintosh VT100 Terminal Emulator I wrote a couple of months ago. I'll >send you either the (Turbo) Pascal source or a 'pseudocode' version of it >if you want. It avoids the nasty skipping past tabs problem that all >move mouse/click/work out diff/send arrow codes systems I've ever seen in >ST and Mac terminal emulators. The problem is that the 'skipping over tab problem' isn't the only problem... >By the way, you might have worked out that I'm working on a VT100 terminal >emulator for Acorn's wonderful new Archimedes machine. I bought one about >10 weeks ago and ARM code is an absolute joy to write. How nice to have enough money for things like that (I can't even afford a printer for my ST). -- ---------------------------------------------------------------------------- UUCP: ...mcvax!cernvax!forty2!poole Simon Poole BITNET: K538915@CZHRZU1A ----------------------------------------------------------------------------