plinio@curtiss.seas.ucla.edu (Plinio Barbeito/) (04/30/91)
As many of you may have heard, last Saturday (Apr. 27) was the date of a conference for which Bob Brodie flew down from Sunnyvale to chat with and and answer the questions of Southern California Area user group representatives (people from as far as San Diego drove up; to everyones' astonishment, one developer made the drive down from Seattle in one day). The format of the conference was: John Tarpinian with the introductions and background, Bob Brodie with a 'state of things' talk, a break, and then a question and answer session. After the conference, one of the dealers, TCN of Glendale, set up developer displays inside the (small and crowded) store and people were able to talk to the company reps and try out (and buy) the software. Most of Brodie's talk concentrated on the dealer network and what Atari is trying to do to strengthen it. He also mentioned management changes and how they are trying to stabilize that by trying to hire only people that already know how to use (and like) Atari computers. He noted his desire to improve the availability of software in the U.S. by sending scouts out at Dusseldorf and Hannover during the big computer trade shows to try to get people to link up with U.S. distributors. Atari is buying the rights to a defunct word processor (wordup? I forget), which they may release under their own label, include bundled on hard disks purchased, etc. Much talk about "retouche", apparently from France. I don't recall exactly what it was for, except that it was graphically oriented and loaded with features that can't be found even on the Mac (sorry about the thinness there). Brodie showed a good sense of humor throughout, and was even willing to make a crack about Leonard. On the whole, it was a pleasant talk, so don't let my criticisms below change that impression. The question and answer question harvested a few interesting tidbits. - Inmos H1 processor will be used in the ATW. I thought I heard something about a change in the operating system, but I didn't catch the name. Mach, I hope. - Factory in Israel is being built because Western Digital would not supply Atari with the controllers it needed, and Atari decided to build its own. Israeli techs reverse engineered the WD1772(?) and the resulting chip can be revved up to higher clock rates to support the higher density floppies. - Tax problems the reason Atari claims as why it has not finalized plans for a facility in Houston, Texas. - Atari looking into a 1 GByte 3.5" disk for the TT (next year). No word on Unix pricing yet (an audience member dogged him until he had to repeat "has not been finalized" a few times). No way to switch between TOS and Unix on the same TT. The techie said that to do so would slow the machine down too much, since they'd have to build an emulator "similar to PC Ditto". This is kind of an unfair comparison, since the processor wouldn't have to be emulated -- all you should need is a toggle for two different ROMS, and a 2-type hard disk partitioning scheme. But they said that the hard disk purchased will determine the (one) operating system. - Long discussion about the memory system in the TT (I left to move my car -- when I was back, they were still talking about it). You can upgrade the ST memory in the TT to up to 10Meg (total of 23 Meg possible, I think). They were treating it as if no-one would ever need that much memory. Ha! I've already used machines with 32 that didn't have enough (a Unix server). There were other things mentioned which I forgot, because I didn't bring a notepad or a tape recorder (some people did), but I think these were the main points of interest for the Conference. Later in the day, at the TCN open house, I was able to ask a GBS representative a couple of questions. He said that the only thing keeping an SST board from being used in a non-Mega ST is the space inside the case. It mounts in a socketed slot where the 68000 usually goes. Those people that already have tower or other large cases should be able to use it. Furthermore, he said there was no word yet on availability. My intuition tells me it should be at least a month before they start shipping -- the rep didn't bring a prototype, so they probably haven't gone to fabrication and are still busy with the (expensive) prototypes. I also had a chance to ask J. Tarpinian (local user group VIP) about ST sales. He said that they were actually picking up in the U.S. due to their popularity with music dealers and customers. In the LA area, these dealers mention Atari in their radio commercials right next to fast-paced talk about Fender Guitars. Music dealers are also advertising in the print media here, with big, full-page ads. Well, if anyone else attended, maybe you can fill in some more details. plin -- ----- ---- --- -- ------ ---- --- -- - - - plinio@seas.ucla.edu I speak for myself, not for my employer. I am a bad ventriloquist, anyway.
kiki@uhunix1.uhcc.Hawaii.Edu (Jack W. Wine) (04/30/91)
First of all, thanks for the interesting report! >plinio@curtiss.seas.ucla.edu (Plinio Barbeito/) writes: >- Inmos H1 processor will be used in the ATW. I thought I heard >something about a change in the operating system, but I didn't catch >the name. Mach, I hope. This is surprising because I thought that someone posted an article stating that the ATW was no longer in production. If the ATW is still alive and will be H1-based, won't the principle market be research-oriented institutions? A detailed description of the H1 was recently posted by an Inmos engineer to comp.sys.transputers, btw. Some specs of the H1 (T9000) are: two million transistors on a 180 sq.-mm die, 50Mhz, 200 native Mips peak, 50-80 native Mips sustained, operates on up to 40 instructions simultaneously. Expected volume production in mid '92. >- Factory in Israel is being built because Western Digital would not >supply Atari with the controllers it needed, and Atari decided to >build its own. Israeli techs reverse engineered the WD1772(?) and the >resulting chip can be revved up to higher clock rates to support the >higher density floppies. This is really puzzling news! If Atari built a factory in Israel, it can't possibly be dedicated to only floppy controller chip production. I may be mistaken, but aren't all those chips becoming extinct because of the prolifer- ation of the SCSI bus in systems and peripherals? Atari might be better off ordering 10,000 or so SCSI floppies and offering them to Mega STe and TT owners for $100 and their 720K drives. They could also off- er older ST systems owners an upgrade which includes a 2.88MB SCSI floppy and a SCSI host adaptor for $200-250. This would standardize all ST owners around the SCSI bus and a relatively-fast, high-capacity floppy drive. I'm not sure, but I think that a SCSI floppy is also able to remap bad sectors of a disk to good sectors so that you probably won't have to throw any disks away. It seems like a short-term and expensive decision, because many people believe the benefits of the SCSI intelligence would cause manufacturers to leave out the floppy controller chip in future systems. >- Atari looking into a 1 GByte 3.5" disk for the TT (next year). No >word on Unix pricing yet... The TT-Unix combo will probably be too expensive for most people and I wish Atari would license OS/68K from Microware and package it with TT and ST lines. If Atari can put together a software/hardware bundle with Microware that includ- CD-I support, this might give TT developers advance info about CD-I, which is supposed to debut this fall. Then they could program for a couple of dozen million CD-I users! >- Long discussion about the memory system in the TT (I left to move >my car -- when I was back, they were still talking about it). You >can upgrade the ST memory in the TT to up to 10Meg (total of 23 Meg >possible, I think). They were treating it as if no-one would >ever need that much memory. Ha! I've already used machines with 32 >that didn't have enough (a Unix server). I don't know about the upgrade options on the TT, but does anyone know if the Mega STe is capable of addressing 16 megabytes? If it supports the VME A24/D16, then that would be true(?) Could someone check in their machines and count the number of pins on the Mega STe MMU and compare it with a ST MMU? The new Atari User magazines listed the specs the the ST(e) Notebooks and it stated that there were two 4MB ramcard slots. Along with the max. 4MB drams and 512KB rom, that adds up to 12.5MB. This means new MMUs (at least for the Notebook) probably support 24-bit addressing. Does the Mega STe have this MMU (and the regular STe) too? Could someone please check? >I also had a chance to ask J. Tarpinian (local user group VIP) about >ST sales. He said that they were actually picking up in the U.S. >due to their popularity with music dealers and customers. In the >LA area, these dealers mention Atari in their radio commercials >right next to fast-paced talk about Fender Guitars. Music dealers are >also advertising in the print media here, with big, full-page ads. I noticed the Atari logo on some ads from dealers that previously sold Amiga and IBM compatibles. It seems that Atari is laying the foundations for a strong and lasting resurgence. Thanks for the good news! Jack
chuck@mrcnext.cso.uiuc.edu (charles bridgeland) (04/30/91)
.......................plinio.......................- Atari looking into a 1 GByte 3.5" disk for the TT (next year). No word on Unix pricing yet (an audience member dogged him until he had to repeat "has not been finalized" a few times). No way to switch between TOS and Unix on the same TT. The techie said that to do so would slow the machine down too much, since they'd have to build an emulator "similar to PC Ditto". This is kind of an unfair comparison, since the processor wouldn't have to be emulated -- all you should need is a toggle for two different ROMS, and a 2-type hard disk partitioning scheme. But they said that the hard disk purchased will determine the (one) operating system. ..................................................................... bob brodie came to our last months users group meeting. I asked him about unix and he started talking about "at&t version 4" (??). he also said there was both some kind of hardware problem and a small problem in the source from at&t. unix XOR tos? big :-( -- ---------------------------------------------------------------------------- chuck bridgeland---anarchoRepublican "one thing about a police state, you can always find the police" l. neil smith chuck@mrcnext.cso.uiuc.edu hire me so I can quit this pit. -----------------------------------------------------------------------------
plinio@babbage.seas.ucla.edu (Plinio Barbeito) (05/01/91)
In article <12733@uhccux.uhcc.Hawaii.Edu> kiki@uhunix1.uhcc.Hawaii.Edu (Jack W. Wine) writes: >First of all, thanks for the interesting report! No prob. >>plinio@curtiss.seas.ucla.edu (Plinio Barbeito/) writes: > >This is surprising because I thought that someone posted an article stating >that the ATW was no longer in production. If the ATW is still alive and will >be H1-based, won't the principle market be research-oriented institutions? As a matter of fact, that's exactly what Brodie mentioned -- that it was intended mainly for people that were going to write their own software anyway. Institutional or Corporate Research. This was the case with Eastman-Kodak. They bought the ATW hardware for a specific purpose and then said "get out of our way!" (software wise). I don't know whether the ATW is in production or not. Probably doesn't matter, because the demand seems to be for only a very few units. >This is really puzzling news! If Atari built a factory in Israel, it can't >possibly be dedicated to only floppy controller chip production. I may be Sorry to have implied that. I don't know if it will be dedicated to that sole purpose. That's all that Bob mentioned about it, though. >Atari might be better off ordering 10,000 or so SCSI floppies and offering them I agree, WD and plants in Israel may go away (Scud? :-) ), but SCSI is here to stay (well, at least SCSI-2 is). However, I think that if the 20.8M 3.5" drives read 720K and 1.44 Meg floppies, Atari should skip the whole lot and offer that instead. If, as you mention, everyone had this standardized quasi-harddisk capability, new software writers wouldn't be pressed to fit everything into a few disks, lowering development time and keeping the cost of good software low. It would raise the cost of small software, until the 21M floppies come down in price, but it would simplify backups without requiring the purchase of a tape drive. Would I go so far as to say, skip the 20.8's and go on to Syquest? I can't stick those things in a shirt pocket, or in a folder, for that matter. IMHO, 5.25" is a step backwards. Also, I haven't seen a Syquest mechanism for less than $585; to add $500 to entry-level machines is really pushing the startup cost. To add $300 and kill two birds with one stone (720K compatibility + real storage) seems the better solution. >It seems like a short-term and expensive decision, because many people believe >the benefits of the SCSI intelligence would cause manufacturers to leave out the >floppy controller chip in future systems. I think Atari has an (unfounded?) fear of disk incompatibility with the current base of software. True, the last thing they need is for the software base to be shrinking. If they maintain the same controller for 720K-2.8M, they probably think it will be easier to solve any compatibility problems. Yes, it seems like building a new plant is more of a risk than using a (possibly?) incompatible 20.8M drive. However, if EC turns into "Fortress Europe", and Atari is caught without a more local plant by 1992, they could see their primary market disappear... Also, does selling duty-free to Europe mean they can get around the 15% Value Added Tax (ouch)? >I noticed the Atari logo on some ads from dealers that previously sold >Amiga and IBM compatibles. It seems that Atari is laying the foundations >for a strong and lasting resurgence. Thanks for the good news! For the most part, it looks like they are making more of the "right" decisions. -- ----- ---- --- -- ------ ---- --- -- - - - plinio@seas.ucla.edu I speak for myself, not for my employer. I am a bad ventriloquist, anyway.
stephen@oahu.cs.ucla.edu (Steve Whitney) (05/02/91)
In article <2623@lee.SEAS.UCLA.EDU> plinio@babbage.seas.ucla.edu (Plinio Barbeito) writes: `In article <12733@uhccux.uhcc.Hawaii.Edu> kiki@uhunix1.uhcc.Hawaii.Edu (Jack W. Wine) writes: `>First of all, thanks for the interesting report! `No prob. ` `>>plinio@curtiss.seas.ucla.edu (Plinio Barbeito/) writes: `>This is really puzzling news! If Atari built a factory in Israel, it can't `>possibly be dedicated to only floppy controller chip production. I may be ` `Sorry to have implied that. I don't know if it will be dedicated to `that sole purpose. That's all that Bob mentioned about it, though. I didn't think he said it would be a factory. Sounded to me like they had hired people to reverse engineer the WD1772 and do a layout for one that would run faster. `>Atari might be better off ordering 10,000 or so SCSI floppies and offering them ` `I agree, WD and plants in Israel may go away (Scud? :-) ), but SCSI is `here to stay (well, at least SCSI-2 is). ` `However, I think that if the 20.8M 3.5" drives read 720K and 1.44 Meg `floppies, Atari should skip the whole lot and offer that instead. `If, as you mention, everyone had this standardized quasi-harddisk `capability, new software writers wouldn't be pressed to fit everything `into a few disks, lowering development time and keeping the cost of good `software low. It would raise the cost of small software, until the `21M floppies come down in price, but it would simplify backups without `requiring the purchase of a tape drive. Would I go so far as to say, `skip the 20.8's and go on to Syquest? I can't stick those things in a `shirt pocket, or in a folder, for that matter. IMHO, 5.25" is a step `backwards. Well, it would do more than that. As Bob mentioned, parts of GEMDOS (BIOS and/or XBIOS?) are closely wed to that controller. If they used a new controller, it would be very difficult to make it appear to be the WD1772, and software that talks to the contorller would also fail. `>It seems like a short-term and expensive decision, because many people believe `>the benefits of the SCSI intelligence would cause manufacturers to leave out the `>floppy controller chip in future systems. ` `I think Atari has an (unfounded?) fear of disk incompatibility with `the current base of software. True, the last thing they need is for `the software base to be shrinking. If they maintain the same controller `for 720K-2.8M, they probably think it will be easier to solve `any compatibility problems. Yes, it seems like building a new plant is `more of a risk than using a (possibly?) incompatible 20.8M drive. `However, if EC turns into "Fortress Europe", and Atari is caught `without a more local plant by 1992, they could see their primary `market disappear... Also, does selling duty-free to Europe mean they `can get around the 15% Value Added Tax (ouch)? ` Another problem is that Western Digital told ATari that there would no longer be a WD1772. What do you put in 1040STEs? SCSI floppies would be awfully expensive for a low-end machine. `-- `----- ---- --- -- ------ ---- --- -- - - - plinio@seas.ucla.edu `I speak for myself, not for my employer. I am a bad ventriloquist, anyway. -- Steve Whitney - UCLA CS Grad Student (())_-_(()) Soon to be working at Silicon Graphics | (* *) | Internet: stephen@cs.ucla.edu UCLA Bruin--> { \_@_/ } GEnie: S.WHITNEY `-----'
plinio@curtiss.seas.ucla.edu (Plinio Barbeito) (05/02/91)
In article <1991May1.180512.2572@cs.ucla.edu> stephen@oahu.cs.ucla.edu (Steve Whitney) writes: >In article <2623@lee.SEAS.UCLA.EDU> plinio@babbage.seas.ucla.edu (Plinio Barbeito) writes: >`In article <12733@uhccux.uhcc.Hawaii.Edu> kiki@uhunix1.uhcc.Hawaii.Edu (Jack W. Wine) writes: ... >`...It would raise the cost of small software, until the >`21M floppies come down in price, but it would simplify backups without >`requiring the purchase of a tape drive. Would I go so far as to say, >`skip the 20.8's and go on to Syquest? I can't stick those things in a >`shirt pocket, or in a folder, for that matter. IMHO, 5.25" is a step >`backwards. > >Well, it would do more than that. As Bob mentioned, parts of GEMDOS (BIOS >and/or XBIOS?) are closely wed to that controller. If they used a new >controller, it would be very difficult to make it appear to be the WD1772, >and software that talks to the contorller would also fail. This would only matter to copy-protected games that use a disk-based copy protection scheme. Because the O/S provides fast and flexible disk routines, the vast majority of software doesn't re-invent the wheel and uses them, even many slick games. I realize that Atari considers the rule-breaking games an important asset in its software base. However, eventually they are going to have to make the jump to higher density media. How large a jump should be made without creating 3 different formats? What format will most packaged software start coming out in? If the jump isn't large enough, companies won't bother to use the larger format, try to sell to a larger base, and just ship a bunch of 720K disks that the user is going to have to swap in and out and that will limit the size of executables. If one jumps all the way out to 21M, things become easy that were once inconvenient or impossible, and a software company might release a product in that format simply because it adds so much to the capabilities compared to other packages coming out in the small format. Let's take the example of a game. High speed sound samples and pictures take up a lot of space. With 21M available, these things become pratical; you can sprinkle them liberally throughout the software, benefitting creativity. Someone might be tempted to release a Mega-game in that format, pulling customers towards the new technology like flies. Moreover the samples and executables wouldn't take 5 minutes to load because: 1) transfer rate would be radically faster, 2) decompression would not be needed, 3) You can use DMA. As long as we stick with 720K, there's only so many high-speed samples you have access to at once. Improvements would be similar with business software, including the fact that users wouldn't have to worry about installation (yet), already giving the software an ease-of-use advantage. >Another problem is that Western Digital told ATari that there would no longer >be a WD1772. What do you put in 1040STEs? SCSI floppies would be awfully >expensive for a low-end machine. Right now, yes. I like the idea of seeing very cheap hardware available. It does wonders to increase the market share. But eventually these people try to do serious things with their machines and will come to the conclusion that one, 800K drive is not enough for their needs and will want to upgrade their hardware. When they do, Atari should be ready for them with a standard. What I am saying is they should be hinting at developers and users that the standard they should upgrade to is 20.8 megs by using the technology in the high-end machines. Even today a $300 20M drive would already be a cost-saving option for the Mega STE. Get rid of the hard disk and floppy and replace it with one unit. In the future, it is not unreasonable to expect that the price of the 20.8 megs will come down to a more affordable level, making them practical even on low-end machines. Lastly, there is the competitive factor. A few people will make the decision on whether to buy or not to buy a computer based on whether they offer better hardware than the competition. If 720K is all that's available, they may be prompted to look at a different system. Cheap clones commonly come with 1.2 or 1.4M drives and hard disks now, making 800K seem less functional and therefore less attractive. In all of this I am assuming that the media cost of the 20.8's will be affordable. If not, then all of this is garbage. 1/2 :-) -- ----- ---- --- -- ------ ---- --- -- - - - plinio@seas.ucla.edu I speak for myself, not for my employer. I am a bad ventriloquist, anyway.
jclark@sdcc6.ucsd.edu (John Clark) (05/02/91)
In article <5571@wucc.waseda.ac.jp> ytsuji@wucc.waseda.ac.jp (Y.Tsuji) writes:
+It's timely that WD1772 was finally discontinued. I also wish Motorola to
+discontinue 68000,68010 and 68012. They are so dated! As to the downward
No, no, not the 68000 (you can have the other two). How about a last
round for the 8080, z80, and anything ending in '86.
--
John Clark
jclark@ucsd.edubill@mwca.UUCP (Bill Sheppard) (05/03/91)
In article <1991May1.180512.2572@cs.ucla.edu> stephen@oahu.cs.ucla.edu (Steve Whitney) writes: `In article <2623@lee.SEAS.UCLA.EDU> plinio@babbage.seas.ucla.edu (Plinio Barbeito) writes: ``In article <12733@uhccux.uhcc.Hawaii.Edu> kiki@uhunix1.uhcc.Hawaii.Edu (Jack W. Wine) writes: ``>>plinio@curtiss.seas.ucla.edu (Plinio Barbeito/) writes: ``>This is really puzzling news! If Atari built a factory in Israel, it can't ``>possibly be dedicated to only floppy controller chip production... `` ``Sorry to have implied that. I don't know if it will be dedicated to ``that sole purpose. That's all that Bob mentioned about it, though. ` `I didn't think he said it would be a factory. Sounded to me like they had `hired people to reverse engineer the WD1772 and do a layout for one that `would run faster. Both the San Jose Mercury News and the Jewish Bulletin reported that Atari and the Israeli government are negotiating (close but nothing signed yet) for Atari to open a manufacturing plant in Israel, with significant incentive (financial) from the Israeli government, who are hoping to start a high-tech business park. The plant would apparently produce most types of Atari hard- ware (though I believe the Singapore plant would remain) and would give (again, I'm not positive) Atari major advantages in dealing with the EC. -- ############################################################################## # Bill Sheppard -- bills@microware.com -- {uunet,sun}!mcrware!mwca!bill # # Microware Systems Corporation --- OS-9: Seven generations beyond OS/2!! # ######Opinions expressed are my own, though you'd be wise to adopt them!######