F18A - 80 Column & Enhanced Graphics Supported Programs

[matthew180]

Benefiting? I suggest you try it first hand before you attempt to apply economics to something like this.

[slinkeey]

I never said that *you* were benefiting from it! Hey, you're the one that mentioned that you keep building them because people keep asking for them. (Supply and Demand) If it was such a crappy product, the market would tell you. You know, vote with your wallet!

 

 

I would guess that you are lucky to even be breaking even.

Edited November 6, 2013 by slinkeey

[+OLD CS1]

Benefiting? I suggest you try it first hand before you attempt to apply economics to something like this.

 

I believe the stab was at Gazoo and not you: you keep building them because of a demand for them. We have Gazoos in every retro community I am a member.

[matthew180]

@slinkeey: Sorry, I misunderstood what you were saying. I would love to quit my job and make retro-hardware full time. I don't know if that is possible though.

Edited November 6, 2013 by matthew180

[slinkeey]

No problem! I am used to it because I always have that problem. Not sure why, but I do.

[TI-Sissy]

Please, just a friendly reminder..

 

This thread is devoted to the F18A and the programs that can utilize it's extended capabilities.

 

I'd rather not see this thread degrade into a "9938 or 9958" opinion debate. Thank you.

Please keep it positive and useful .

 

not working out too well i see

 

 

Sweet! At the rate new stuff is coming out for the F18A, Matthew might have to start an assembly line! I noticed of the last batch he put up a couple of days ago only two are left. People are really snatching this things up. I doubt the last two make it very long before that "out of stock" sign goes up again.

 

i want one too. i will have to save for a new monitor too. 3 or 4 months? the way you gush over it and the others buyers seem to like it too, i'm sold

 

 

Maybe you should get a little more educated about the graphics programs written for the TI & Geneve that use the bitmap graphics of the 9938/58. And you have admitted to knowing next to nothing about the Geneve. Your reply is your opinion, which has done absolutely nothing to convince me to change my observation. 'Deficiencies' is the correct word - lack of vram and inability to support all the 9938/58 registers being the 2 main deficiencies of the F18A.

 

I'm sure it's a neat little device, and many people will use it and enjoy it, but it's not for me - I'm too used to the 9938's in my TI & Geneve to take a step back. In fact, I was considering buying one just to check it out, but have nothing to plug it in to.

 

Gazoo

 

 

why r u being so nasty? the only people i see poo pooing it are the guys who don't have one. everyone who has one seems happy.

[slinkeey]

Sissy,

 

You can get a monitor pretty cheap or free.. People like the wide screens and are dumping the standard 4:3 format monitors.

 

I actually have an extra.

Edited November 6, 2013 by slinkeey

[+OLD CS1]

You can get a monitor pretty cheap for free.. People like the wide screens and are dumping the standard 4:3 format monitors.

 

I actually have an extra.

 

I try to hold onto my 4:3 stuff as much as possible. Some of those old Dell monitors which tend to go out can be fixed with a simple solder re-flow or transistor replacement.

[+InsaneMultitasker]

Ok.. learned a couple things,

1: don't use dm2k to load it..

2: if you rename the game files to game1 game2 that works but the screen is hashy(see attachment)

 

Greg

 

 

 

I PM'd you a modified version to try on the real hardware. Not sure if you get those notifications or not

[+arcadeshopper]

I PM'd you a modified version to try on the real hardware. Not sure if you get those notifications or not

 

I did just was unconcious will look tonight and let you know

[+Lee Stewart]

With the text-modes block loaded (currently starts at block 30 of FBLOCKS), fbForth supports 80-column text mode ( TEXT80 ) when the F18A is present on Classic99 and the real iron. It supports 80-column text mode on MESS with the SNUG 80-column card present and I presume the same situation obtains on the real iron, but haven't tested it—I don't have such a card. The 40/80-column editor (currently starts at block 13) allows editing an entire 16-line-X-64-column block in TEXT80 mode.

 

...lee

[matthew180]

And 80 columns in a 9938 or 9958 does not look like 80 columns on the F18, as it used the extra RAM to get 80 columns in memory.

What do you mean by "does not look like"? I implemented the 80-column mode on the F18A based on the 9938 datasheet. On the monitor it "looks the same" as the 9938, and in the first 16K of VRAM it "looks the same" as far as the memory utilization and register setup goes.

 

Does the F18 use the extra RAM to store the actual VDP tables like these chips did?

The "extra" RAM, i.e. the 2K of RAM above the original 16K of VRAM is accessible to the GPU only. So no, you cannot locate any name tables, pattern tables, sprite tables, or color tables in the extra 2K of RAM.

 

If you stay in the original 16K of VRAM, the F18A can work with 80-columns just like the 9938.

 

I am asking as F18 is on Classic99 so I can take advantage of it, want me to warm up to it right?

Do *I* want you to warm up to the F18A? It would certainly be cool to see F18A support in RXB, or any other programs you write, but do whatever it is that makes you happy. If you want to check out the F18A, great, it will be cool to see what you come up with. If not, well, that's fine too.

[+arcadeshopper]

Sissy,

 

You can get a monitor pretty cheap or free.. People like the wide screens and are dumping the standard 4:3 format monitors.

 

I actually have an extra.

 

I can pick up used 7-19" 4:3 LCD's at the local computer recycler for 35-40$ in quantity.. happy to ship if there is demand for costs incurred. I put two on my Amiga's that had video cards that could handle them.. no brainer.. also there's the Samsung 910mp found periodically on ebay for about $100 that's a tv,monitor and has SCART capable of handling the 15hz analog RGB people are fond of, so good for both old systems AND the f18a I just picked up one for my Geneve, now just have to make a cable..

 

Greg

[Gazoo]

 

why r u being so nasty? the only people i see poo pooing it are the guys who don't have one. everyone who has one seems happy.

 

Not nasty, I started out by mentioning that had the 9938/58 registers been implemented correctly and enough ram been included on the F18A, all these patches to long established and stable software would not be necessary. Just making an observation. The ranting on my simple observation is what I would consider nasty.

 

Gazoo

[TheMole]

 

Not nasty, I started out by mentioning that had the 9938/58 registers been implemented correctly and enough ram been included on the F18A, all these patches to long established and stable software would not be necessary. Just making an observation. The ranting on my simple observation is what I would consider nasty.

 

Gazoo

 

There were two things wrong with your initial comment that made it come across really nasty:

1. You immediately assumed that there could not have been a bug in BOOT because it works on the 9938. Clearly, there could be a bug in the 9938's implementation versus the specs and documentation. It is perfectly feasible that Matthew's implementation is actually more correct than Yamaha's. Not saying this is the case 'cause I simply can't know without looking at the patch, but neither can you so it would behoove you to give Kevan's statement the benefit of the doubt if you don't plan on making the effort to prove him wrong.
2. We know by now that the F18A is not for you, and that is fine. But instead of spending your breath reiterating your dislike in these F18A specific threads you should probably start your own 9938/9958 threads where you can talk about the things you like about it instead of talking about what you dislike about the F18A. Maybe you can gain some interest and someone will be triggered think of ways to make 9938/9958 boards more widely available, choice is good, but there's no point to take an anti-F18A stance in F18A topics.

I promise you that if you start threads about the 9938 I will not keep talking about how the programming models are awkward and most of the extensions not well designed for my personal prefered type of games. I promise I will not keep shoving the F18A in everyone's face as the clearly superior alternative in those threads.

 

 

Same goes for Rich and YAPP or whatever non F18A compatible program you want to talk about. If you think it's such a cool thing, why do you think it doesn't deserve its own thread? I would love to hear more about it in the right context. I would love to see art made in it, I would love some screenshots or maybe some youtube clips. Why the hell not, it sounds like an interesting program. I am being very sincere about this, my google searches yielded nothing interesting except a fuzzy scan of the manual...

But it's just not appropriate in this context, this specific thread is about programs that use the features of the F18A, not about programs that use features the F18A doesn't have (unless you want to propose a way to port it over).

[Willsy]

Putting it in binary terms:

 

 

do
 
loop until(understood(f18a != v99x8))

[marc.hull]

I think Tony's only objection was to Kevan calling a piece programming buggy because it wouldn't work on the F18 without modification. Semantics maybe but that seems his intent.

 

So Kevan, What made BOOT80 "buggy" and how did you change it so it would run on an F18 ?

Edited November 7, 2013 by marc.hull

[Omega-TI]

I think Tony's only objection was to Kevan calling a piece programming buggy because it wouldn't work on the F18 without modification. Semantics maybe but that seems his intent.

 

So Kevan, What made BOOT80 "buggy" and how did you change it so it would run on an F18 ?

 

 

Me? I didn't do anything, Tursi is the one who fixed/patched it. Tursi said it was a "one-byte" patch to zero the unused bit in VDP register 2.

You can read a more in-depth explanation from Tursi in his post on the Yahoo site here:

http://groups.yahoo.com/neo/groups/ti99-4a/conversations/messages/84695

 

Well, Tursi said it was a bug in BOOT, he fixed the problem, and explained it pretty darn well!

Besides, who am I to question Tursi? The guy is a freaking genius! I figure he's correct. And I never argue with results!

[marc.hull]

Me? I didn't do anything, Tursi is the one who fixed/patched it. Tursi said it was a "one-byte" patch to zero the unused bit in VDP register 2.

You can read a more in-depth explanation from Tursi in his post on the Yahoo site here:

http://groups.yahoo.com/neo/groups/ti99-4a/conversations/messages/84695

 

Well, Tursi said it was a bug in BOOT, he fixed the problem, and explained it pretty darn well!

Besides, who am I to question Tursi? The guy is a freaking genius! I figure he's correct. And I never argue with results!

I followed your link but I did not see Mike calling the software buggy. Only that it needed a patch to make it F18 compatible. In fact here is a quote from that thread that pretty much sums things up.....

 

"Nobody should be surprised that software written for the 9938 needs a little work to run comparable on the F18A -- the F18A is not a 9938 clone, it's a 9918A clone with some extra features....."

 

What am I missing here ?

[Omega-TI]

I followed your link but I did not see Mike calling the software buggy. Only that it needed a patch to make it F18 compatible. In fact here is a quote from that thread that pretty much sums things up.....

 

"Nobody should be surprised that software written for the 9938 needs a little work to run comparable on the F18A -- the F18A is not a 9938 clone, it's a 9918A clone with some extra features....."

 

What am I missing here ?

 

Possibly when he wrote, "I can't follow why the author set the bits that way, probably had some bad documentation and just accepted when it worked." I interpreted that to mean bug, others may see it differently, but none of this really means anything, we all see things through our own filters.

 

Now I think the "TheMole had an excellent suggestion, all future 9938 and 9958 discussions should reside in a different thread. Thanks.

[marc.hull]

Possibly when he wrote, "I can't follow why the author set the bits that way, probably had some bad documentation and just accepted when it worked." I interpreted that to mean bug, others may see it differently, but none of this really means anything, we all see things through our own filters.

 

Now I think the "TheMole had an excellent suggestion, all future 9938 and 9958 discussions should reside in a different thread. Thanks.

I guess I glossed over that. Thanks for pointing it out (in large letters no less.)

 

What evidently was happening (according to Mike's assumptions and plenty of assumption on my part) was that some bits normally unused and invisible by the 38/58 have been re-purposed in the F18? So when the hardware platform changed the software caused the F18 to go into a bad mode that normally wouldn't happen with the 38/58? If so that's not really a "bug" but more of an unintended consequence of changing hardware behaviors without patching the software. But like you said everyone has their perception. I guess Tony's was from the vantage of an old school 9900/9995 9918/38/58 programmer and he got a bit defensive at your statement that the old stuff was and had been buggy for decades.

 

Does the 9938/58 ban include you asking for software to be ported over from the... well...... 9938/58 ?

[RXB]

What do you mean by "does not look like"? I implemented the 80-column mode on the F18A based on the 9938 datasheet. On the monitor it "looks the same" as the 9938, and in the first 16K of VRAM it "looks the same" as far as the memory utilization and register setup goes.

 

 

The "extra" RAM, i.e. the 2K of RAM above the original 16K of VRAM is accessible to the GPU only. So no, you cannot locate any name tables, pattern tables, sprite tables, or color tables in the extra 2K of RAM.

 

If you stay in the original 16K of VRAM, the F18A can work with 80-columns just like the 9938.

 

 

Do *I* want you to warm up to the F18A? It would certainly be cool to see F18A support in RXB, or any other programs you write, but do whatever it is that makes you happy. If you want to check out the F18A, great, it will be cool to see what you come up with. If not, well, that's fine too.

 

Cool so it looks like the 9938/9958 after all? I assumed wrongly that it was like squished characters like a most other 80 column hacks.

 

So not possible to use the previous popular programs like FW 80 or X80 or First Draft 80 but new program hacks are out there now.

 

RXB would never be able to use 40 Column mode (been there and tried that) but the REA (Rich Editor Assembler) would look great with a 80 column display.

 

Will put a hacked version of REA for F18 on the list of things to do.

[+arcadeshopper]

 

I did just was unconcious will look tonight and let you know

 

results were the same, we're taking our alpha testing convo to pm and report once it's fixed

[matthew180]

VR2 is the name table base address. The 9918A and F18A only use the low 4-bits, while the 9938 uses the low 7-bits to support the expanded VRAM. The F18A will ignore all but the low 4-bits, so I don't see how setting the 4 MSbits would affect anything? I'll have to get some details from Tursi. Unless the original code was expecting to place the name table in VRAM over 16K?

 

VR2:
       0  1   2   3   4   5   6   7
       -----------------------------
9918A  0  0   0   0  A13 A12 A11 A10 (A13..A0)
F18A   X  X   X   X  A13 A12 A11 A10 (A13..A0)
9938   0 A16 A15 A14 A13 A12 A11 A10 (A16..A0)

A0 = LSb in VDP Address, numbered here opposite TI numbering for clarity
X = don't care, totally ignored

As long as the 9938 code stays in the original 16K and sets the 4 MBbits of VR2 to zero, it will behave like the 9918A/F18A as far as the name table location goes. Setting VR2 bits 1..3 to something other than zero on a 9938 will move the name table into memory over 16K.

 

The F18A follows the 9938 register use exactly for the original eight VDP Registers, other than where the 9938 expanded the table location bits since the F18A does not have the expanded VRAM. I also tried to follow the 9938/58 register use where it made sense so that 9938 code might run correctly in cases where F18A and 9938 functionality overlap.

Edited November 8, 2013 by matthew180

[+InsaneMultitasker]

VR2 is the name table base address. The 9918A and F18A only use the low 4-bits, while the 9938 uses the low 7-bits to support the expanded VRAM. The F18A will ignore all but the low 4-bits, so I don't see how setting the 4 MSbits would affect anything? I'll have to get some details from Tursi. Unless the original code was expecting to place the name table in VRAM over 16K?

 

The 80 column Infocom Interpreter does something similar; the pattern table is moved to bank 1 (0x4100) because there is not enough contiguous VRAM space left for the patterns once the name table is expanded to accomodate the 80 column display. I moved the pattern table into a portion of the disk buffer area and patched the display routine to add a needed character offset. Unfortunately, either the program uses more than one file simultaneously or I am misinterpreting file buffer usage.

 

Other 'hacked' programs may use similar tricks which of course limits the ability to "enhance" them for F18A use. I don't have an F18A but needed a challenge to get me through this week. That, and I've always wanted to disassemble the infocom interpreter and now I had an excuse