Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!mnetor!uunet!husc6!rutgers!sunybcs!bingvaxu!leah!uwmcsd1!marque!ddsw1!gryphon!richard From: richard@gryphon.CTS.COM (Richard Sexton) Newsgroups: comp.sys.amiga Subject: Re: Tektronix 41xx (was: Re: Diga! review) Message-ID: <1306@gryphon.CTS.COM> Date: Sun, 23-Aug-87 03:39:58 EDT Article-I.D.: gryphon.1306 Posted: Sun Aug 23 03:39:58 1987 Date-Received: Sun, 23-Aug-87 22:25:56 EDT References: <1052@vu-vlsi.UUCP> <3621@cit-vax.Caltech.Edu> <1245@gryphon.CTS.COM> <260@uscacsc.UUCP> Reply-To: richard@gryphon.CTS.COM (Richard Sexton) Distribution: world Organization: Trailing Edge Technology, Redondo Beach, CA Lines: 159 Keywords: Tektronix In article <260@uscacsc.UUCP> papa@pollux.UUCP (Marco Papa) writes: > >Funny that you think of the 4107 (or 4207) a low-end graphics terminal: >it is priced at more than $2000! $2395. But whose counting. Ours cost a *lot* more than this, we've had it a while. > In reality, the low end is the Tek 4105 >which gives you a little less resolution (480x360), 3 bitplanes for graphics >and 3 bitplanes for text, minus support of Surfaces and Zoom/Pan (which >is included in the Tek 4107). The Tek 4105 can be COMPLETELY emulated Aww, nuts, I forgot about this one. We have all these Tek manuals at work, and the 4107 is the lowest we have. Oops. Dont know why I forgot, I used a C.Itoh clone of this terminal once. >Again, there can be various levels of emulation. Most programs do not use >PIXEL COPY for example, since it is non-portable accross different Tek >terminals, and its use is discouraged in the Tek manuals themselves. Customer uses this feature. Had to include it. Unless I'm mistaken its on all terminals made after the 4107 (ie 412x series). Good luck to Tek for discouraging it. People are using it. Shoot, Tek still includes commands like 'SET_4014_ALPHATEXT_SIZE' to let old 4014 code work properlt on its high end terminals. Blitting is so important these days, that I'd venture we'll see PIXEL_COPY on every subsequent Tek terminal. This was a fun one to implement, much moreso than DELETE_SEGMENT or prompt mode or something. It was kinda neat playing detective figuring out that the terminal was doing with all these wierd permutations and combinations. Interesting thing, but, as you point out, most comtemporary hardware used to emulate Tek terminals often has MUCH higher performence than the original Tek. I have a pixel copy test which takes a good 15 minutes to do on a 4107, but runs in less than a minute under emulation on an apollo DN660. I've programmed the same sequence on the Amiga using 'The Director' from the Right Answers Group, and its still faster yet. About 2 to 3 times faster. Of course the poor old 4107 has an 8086 with no fancy graphics hardware. >>much detail, and you end up having to write a program to try it on >>a real Tek terminal to see what it REALLY does. > >I disagree. We have implemented our Tek4014 just from the manuals and >a few test plots. Well, sure, the 4014 is pretty straight forward. Its the little side effects of the more complex commands sets that arn't documented that get you. I'd hate to try to knock off a 4129 without actually having one. *Much* more complex. > And apart from some "undocumented" features, everything ^^^^^^^^^^^^ ^^^^^^^^ | | | +------ I'm not a bug I'm a ... | +-- Yup, this is what I'm yakking about. >I have found both the Tek4105 and the Tek4107 manuals very complete and >thorough. Yes, I'll have to admit,the 4107 (which is the newest of all mine) is the best of all the tek manuals I have. They have been steadidly been getting better over the years. >I ported their FORTRAN examples to C in 30 min, and they >worked right the first time. I'm glad. One of the examples I tried was : DEFINE_SEGMENT #1 as some shape. Position at (0, 0). SET_SEGMENET_POSITION to (1000, 0) (This 'moves' the shape to (1000, 0)) Fine, that worked. When I changed the example to for (i = 0; i < 10; i) { SET_SEGMENT_POSITION (i*400, i*400) } it did not move the segment across the screen in a diagonal, rather, it ignored the Y component. Teks official response was: "Well, yes it should work, and indeed does on our 4129 here. We dont have a 4107, and besides you shouldnt do that, although technically its legal, you are supposed to do that other ways". Amusing. >What one needs is a good set of plot >examples from some "commercial" packages (we got DISSPLA, ANVIL, GDS, >TEMPLATE, stuff from Precision Visuals, and some others), to see >how they actually use the command set. I dont have any tek output from them, the ones I have are pretty funky. The one I'm working with now is in Italian, has lots of text and seems to go on for about 30K doing: ENTER_ALPHA_MODE ENTER_VECTOR_MODE SET_4014_ALPHATEXT_SIZE 1 SET_4014_ALPHATEXT_SIZE 0 in the middle. Brain damaged, >>I can see no real solution to this 'alphatext overlay' problem > >Wrong! Here the Amiga has a real adavantage over the IBM PC >emulations. The "great" people that designed the Amiga put in a >feature whose purpose was (and is) maily for games. It is called >"Dual-Playfield". One of the dual-playfield modes allows two rasters Yeah, your right, you can do a decent 4105 on an amiga with this. The concession of having 1 less bitplane for alphatext overlay is the least offensive I've seen. Having fewer text colors isn't going to kill anybody, compared to placing the alphatext ONTO the the graphics planes (and then scrolling it) (apollo) or seperate window (sun). >>So, sad but true, the 4014 is probably the only Tek terminal a >>'bare' amiga can do. > >Again I disagree. FULL emualtion of TEK 4105 is real and it is >being done, and CLOSE emulation of TEK 4107 is also possible. Agreed. Although the 4107 may not be close enough for some customers. Some of these swine expect 100% emulation, Now, if somebody were to come out with a 12 plane 1280 x 1024 video card, we could really get fancy. 4125, (basic graphics), 4126 (stereo 3D) 4128 (3D wireframe) and 4129 (Shaded 2D, light sources etc.) Note that the 4129 is 15 or 30K. >>Plus, what do you do about the bugs in the tek terminals ? Do you emulate >>them or fix em. :-) > >You emulate them, since most commercial packages assume and use them. If you emulate the bug where it hangs up the modem when you clear a screen where DABUFFER = 1000 and hardware handshaking is selected, I'll break your kneecaps. :-) Also, I see no point to emulate the SET_SEGMENT_POSITION bug. In fact I can't think or any that are really usefull. Perhaps you could Email me some? >Marco Papa 3175 S. Hoover St., Ste. 275 (213)669-1497 -- Richard Sexton INTERNET: richard@gryphon.CTS.COM UUCP: {akgua, hplabs!hp-sdd, sdcsvax, ihnp4, nosc}!crash!gryphon!richard "It's too dark to put the key in my ignition..."