Historical Interviews with TI employees

[kl99]

Herman Schuurman
by Klaus Lukaschek
Interview taken December 2015

 

Herman Schuurman had a 36 years career at Texas Instruments, from November 1977 to his retirement in 2013. In March 1978 he got promoted to be Lead Programmer for the Consumer Products Group in Lubbock.

The description of his work is taken from LinkedIn for that designation:

Software design for advanced personal computer products. Design and implementation of Text to Speech system based on TMS5200 speech synthesizer; TI 99/4A mini memory development system; I/O section of 99/4 Home Computer; I/O section of BASIC interpreter; system software for various peripheral devices.

[https://www.linkedin.com/in/herman-schuurman-60584b9/]

 

Q) What was it like to work for TI in the Consumer Products Group?

 

It was a lot of fun.  Lubbock is a relatively small community (around 180,000 when I lived there), so we had a tight-knit team there that also got together outside of work. Lubbock itself is desert-like – dry heat in the summer and cold in the winter. It’s part of the south plains; flat as a pancake, with no hills around to speak of.  The names of surrounding towns reflect this: Levelland, Plainview, Shallowater, etc…

I was hired in for the Personal Computer Division in Lubbock, but I worked for the Consumer Products Group in Almelo, The Netherlands while my American work permit came through the system.  In March 1978 I finally moved to Lubbock, having never actually seen the place.

 

Q) Can you describe your relationship to TI as an employee? You almost worked your whole life there.

 

Until recently, TI had a lot of different subdivisions.  This allowed me to work from consumer to industrial systems to research, etc… Lately TI has been concentrating more on analog, so I guess it would be more difficult to stick around for your entire career and still have a variety of assignments.

 

Q) How did it came that you left the Consumer Products Group at TI in 1981?

 

I joined the Dallas-based group that Don Bynum originally came from, the Corporate Engineering Center. By the time I left, most of the system design was over, and the team was moving to application design.

 

Q) Can you describe in detail your involvement with the TMS5200 speech synthesizer?

 

The original design for the TMS5100 was done by Larry Brantingham, Paul Breedlove, Richard Wiggins, and Gene Frantz. Gene was heading up the speech group (home of the Speak & Spell) when I joined TI in Lubbock, and Larry moved to TI France (Nice) soon after.  I eventually wound up in a group with Richard Wiggins when I joined the Corporate Engineering Center in Dallas.

The second generation of the TMS5100, the TMS5200, was used to power the speech peripheral. My involvement with the speech synthesizer was to create the text translation and allophone stringing software in coordination with Kathy Goudie (who worked for Gene), who was responsible for creating the allophone (sound) set and the translation rules.  The linked article by Sharon Crook is basically a rehash of the internal documentation on the text to speech software.

 

Q) All TI-99 Speech Synthesizers have the door that was meant for inserting Speech modules, early units even have an interface for such modules. However no modules were released. Do you have an insight on this? Can you enlighten us with a story about how this was planned and later abonded?

 

The speech module came with 200-odd canned speech phrases that could be used in software modules.  There was a provision for phrase ROMs to be added later to expand the available vocabulary, but the introduction of the text to speech capability made that a moot point.

 

Q) Can you describe your involvement in the TI-99/4 project?

 

The Home Computer (99/4) project started about a year before I joined the team in Lubbock.  I believe the original promotors of the project were Granville Ott and Len Donohoe.  I was originally hired to work on the SR-70, a small scientific computer, but by the time I landed in Lubbock, that project had been moved to the Data Systems Group in Austin, and I was put to work on the SR-62, a small self-contained computer that shared most of its software with the Home Computer. In addition to the Home Computer stuff, the SR-62 had a small built-in monitor and a thermal printer. When the Home Computer eventually fell behind schedule, the entire SR-62 team was moved over to complete the 99/4.

Since my background was in operating system design, I worked on a lot of I/O related stuff such as the audio cassette, thermal printer, etc.  I also was responsible for the I/O section of the BASIC interpreter, including formatted I/O, etc... One of the more complex peripherals was the floppy drive. Bill Nale and I split that design, with Bill responsible for the hardware and the low level software, while I took the file system design and implementation.  This was the only time I remember having contact with anyone from Microsoft, even though a lot of 99/4 websites seem to think that Microsoft was responsible for a lot of the software on the 99/4.  We had Bob Greenberg come out once to validate the file system design (there were no design changes).

 

Q) The TI Dimension 4 almost looks like the TI-99/4 and is from 1978/1979. Do you know anything about it?

Link for Schuurman to the dimension4 atariage Thread

 

It sure looks like an early 99/4; I don’t remember the Dimension 4 name, but it may be an early marketing name for the 99/4.  It was definitely not the Z80-based version, since that looked more like a high-end stereo component, including the wooden side panels.  Compared to the 99/4 it was extremely fast, since the video was memory mapped (and you had a speedy processor).

Before I arrived in Lubbock, there was some work done on a native GPL chip, but by 1978 that had been replaced with an 8-bit TMS9985 based design.  Unfortunately, that chip never ran correctly, so we had to eventually fit a 16-bit TMS9900 into an 8-bit design.  If you like to see more info on the 9985 design debacle, check out Karl Guttag’s page at http://www.kguttag.com/2013/08/10/if-you-havent-tested-it-it-doesnt-work/.

 

Q) Do you still have a TI-99? If yes, can you describe what you have kept. When was the last time you used it?

 

A few years ago I gave all my 99/4a related hardware and software to Joe Zbiciak, since I hadn’t touched it for quite a while. Joe is more into legacy systems, so he had a better use for it.  I kept the Panasonic monitor, but it failed a few weeks ago, so now I’m totally out of 99/4a related stuff.

 

Q) Was your Text to Speech system reused for other Speech products from TI or other companies?

 

Not to my knowledge. A remember Ute Marcotte was working on a German version of the text-to-speech rules/allophone set, but I don’t know if that ever made it out. Most of the later interest seemed to be in speaker independent speech recognition, which I worked on several years later in the Telecom Systems group.

 

Q) Having some relationship with Speech, what is your opinion on Apple's Siri?

 

Compared to the early work we did on speaker independent speech recognition, both Siri and Cortana are phenomenal. Of course the amount of CPU power and available memory space are also incomparable. I use the speech recognition on the iPhone quite a bit, and it is really good.

 

Q) Does your name have an origin in the Netherlands or some close-by country? Can you enlighten us?

 

I was born in the Netherlands and spend the first 24 years of my life there. I graduated from the Technische Hogeschool Twente (now known as the University of Twente) in Enschede, the Netherlands. After graduation I joined TI and moved to the USA.  Although I graduated in Electrical Engineering, a lot of my background was in embedded systems and Operating Systems design/implementation, which is why I was hired by TI.

 

Q) Do you know how it came that the command to load a program is called "OLD" on the TI?

 

It probably came from the original Dartmouth BASIC, where OLD was used to retrieve a program from storage, and NEW to start a new program.

 

Q) Bill Cosby was the front face for the TI-99 to the public, was it a person the employees looked up to as well?

 

Advertising was all handled by the marketing team, but Bill Cosby was very popular in those days, both from his earlier I Spy series, and through the Fat Albert and the Cosby Kids show. In addition, the fact that he had a Doctor of Education degree didn’t hurt.

 

Q) What was the role of Don Bynum you worked with?

 

Don was working on a redesign of the 99/4 while in the Corporate Engineering Center in TI Dallas (the Ranger).  In late 1980 he moved to Lubbock to take over the home computer group from Pete Bonfield (who moved on to become Chairman and Managing Director of ICL in England). He drove the 99/4a and peripheral box efforts, and later the 99/2 and 99/8.  I never actually saw it, but he was rumored to have an old piano cabinet with the 99/4a peripheral chain taking up the entire cabinet.

A few months after the introduction of the 99/4a I moved from the Lubbock team to Bynum’s old group, the Corporate Engineering Center in Dallas.

 

Q) Was it clear from the beginning that there will be some Text to Speech software or was it some brilliant idea by an employee?

 

I don’t know who came up with the text-to-speech idea with the speech group, and Kathy can’t remember either.  I’ll check with Gene Frantz to see if he still remembers. Keep in mind that it was not a given that text-to-speech would actually work acceptably in the 9900 until we actually implemented it.

 

Q) So we could say, that your work on the Text to Speech made the speech modules for the Speech synthesizer irrelevant, right?

 

Yes.

 

Q) On what scientific work was your Text to Speech software based on? Was there any linguist or phonetic scientists involved?

 

TI already had an active speech team located in the Lubbock (one of their best known products is probably the Speak N Spell).  The person I worked with, Kathy Goudie, has a Ph.D. in Linguistics, and the allophones used in the text-to-speech software were created from an extensive speech database that had already been developed within the speech team.

 

Q) I couldn't find much information on the SR-62 and the SR-70 computers, have they ever been released? What have the specs of the SR-62 been? Since you said it should have shared its software with the 99/4, this might be interesting for our people to know.

 

Not surprising, since these are internal designators, and neither product ultimately made it to market.  The SR-62 was essentially a self-contained home computer with a built-in monitor and a thermal printer.

 

Q) Do you know how much Microsoft or Bill Gates was involved in the TI Basic / System Rom of the 99/4?

 

Microsoft was not involved with the 99/4 development.  They (in the form of Bob Greenberg) were contracted to develop BASIC for the SR-70 (which is also sometimes referred to as the 99/7), but the BASIC for the 99/4 was developed in-house.

 

Q) Why was the native GPL chip replaced with the TMS9985? How far was the GPL chip developed?

 

Although it was before my arrival in Lubbock, the GPL chip was supposed to be developed for an external customer. When that customer dropped out, the GPL chip was also dropped, and was replaced by the TMS9985.

 

Q) Accessing peripheral devices through a DSR interface allows us to connect modern devices to the TI without modifying the System Rom. Do you know who came up with the idea on this? Did TI had that before the 99/4 already?

 

It’s been too long to remember the exact details, but it was probably worked out between Bill Nale and me.  Bill would have handled the level 1 features (hardware communications, sector read/write, disk formatting), and I would have handled the file related features. Since the hardware was developed concurrently with the software, it allowed me to work on the file system code by simulating the low-level routines on a TI-990 minicomputer, using a large file on the minicomputer hard drive to simulate a floppy disk.

Keep in mind that I already developed a similar system for an Intellec-80 (Intel 8080 based) system using 8” floppies as part of my thesis.

 

Q) How did you feel when you heard that TI will leave the home computer market after being involved in setting it all up?

 

Sad, although it was probably inevitable due to the financial losses. I’m still glad I had a chance to work on it though.

 

Q) What was the policy of TI with hardware/software/documentation/schematics on canceled projects? It seems a few lucky employees got a TI-99/8 when TI left the market.

 

Typically you have to get official permission to legally get any of this stuff.  That said, it is entirely possible that management at that time gave away some memorabilia.

 

Q) How do you feel when you hear that those machines are still running after all those years and there are some people still doing stuff with them?

 

Absolutely amazed and delighted.  I would have never guessed the 99/4 would survive for over 35 years, especially with the typical rapid turn-over in the computer age.

 

Best Regards,

Herman Schuurman

[kl99]

Rick Payne
by Klaus Lukaschek
Interview taken December 2015

 

Rick Payne was born in Swindon, England while his father was in the US Air Force. His LinkedIn Profile is no longer available, so sorry for the lack of his work details. He did work in the Consumer Products Division of TI in the early 1980's - 1980 to 1983.  He was in Quality Control working on the TI 99/4A and  the TI 99/8 during that time span.  He got employed with TI again in 2009, working in semiconductors.

 

Q) How did you get in touch with Texas Instruments?

 

It was 1976 and I was in Technical School in Little Rock, Arkansas.
Recruiters from TI came to my school and gave a test to the top 5 students.
They were very interested in grade point standing and also class attendance history.
TI was the number 1 company to recruit from the school at that time.
I passed the test and I was offered a position as a Technician in TI’s calculator division in Lubbock, Texas for 4.44 USD / hour.

Q) What was your first day like?

 

This was October 15, 1976
Very exciting and a little nervous.  Nothing like I had experienced before in my life.  The TI calculator was pretty much known by every culture on the planet.
In 1976, TI Lubbock was the hub of TI’s calculator division and manufactured millions a month.

 

Q) Can you describe in detail what you did for Quality Control Engineering?

 

Only the 99/4A console was manufactured in Lubbock.
The plug in GROMS (game modules), expansion box / boards, speech synthesizer, modem, joysticks, etc.. were manufactured at other TI sites.  The original power supply (internal) and the transformer were manufactured by 3rd party.
Epson manufactured the dot matrix printer.   
I had a few job responsibilities in Quality Control -  incoming QC (testing incoming electronic components and circuit boards), submissions to UL, ULC, SA and CE  for their safety certification, manufacturing line audits, console failure analysis – from burn-in testing (line failures) and field failures (after sales).  I also did the failure testing on the power transformer recall that you ask about.  More detail on that later.

 

Q) At which stage of development was the 99/4A when you started at TI?

 

In the summer of 1980  when I transferred from the semiconductor FAB at TI Lubbock to Home Computers, the 99/4 was being manufactured.  The 99/4A had not been released from Product Engineering until the summer of 1981.
We were in the process of building and testing 99/4A prototypes, Beta units in the summer,  fall and winter of 1980 /81.

 

Q) Can you describe in detail your involvement with the 99/4A?

 

Prior to manufacturing release, I worked with Product Engineering to produce a reliable, manufacturable console. I helped to write the quality inspection procedures for each line operation.  This would be the specification that the QC line inspectors would use to test and determine a pass / fail.  
I also spent a month sitting on the assembly line to repair consoles at certain stages.  This included the initial power up after wave solder up to final test and inspection, just before packaging.  

 

Q) When the 99/4A did go into production, did your group felt it was ready, it was tested enough?

 

The summer of 1981.  Yes, All groups signed off on the release to begin manufacturing runs.

 

Q) What is your favorite memory from working in the TI Consumer Products Division?

 

Many memories – Bill Cosby, who was the TI spokesman on TV, visited the manufacturing lines.  The popularity of the 99/4A was immense with about 30% market share at the peak.  Also, I remember working a lot of over time when the power transformer recall happened – testing, devising and implementing a fix.
I learned BASIC programming at this time which I have used a lot in my career.

 

Q) What was the mood like in the office? Did you take a break to play TI-99 games sometimes?

 

The mood was good most of the time - Yes, frequently – TI Invaders and Parsec

 

Q) What tools did you use to test the quality of the products?

 

Most of the tools were equipment built by TI and that plugged into the IO port.
Other testing after the console was assembled (final test) was manually performed which required the inspector to pay attention to the monitor for visual and audible function/ quality.

 

Q) How did you test a product? Can you describe that on an example you still remember?

 

Since the output was video and audio, most of the testing was for the tester (human) to use a special GROM that had different test steps.  They would view the monitor and listen to the audio – testing the video RAM, video quality, audio tones, joystick interface, etc...

 

Q) Have you been involved in testing software as well? Was that even part of Quality Control Engineering?

 

No, I was not involved in software testing for any of the TI Home Computer products.  There was a speech synthesizer lab in Lubbock and I believe that the software group for the GROMS (graphics read only memory) modules was in TI Austin, Texas.  

 

Q) Did Quality Control Engineering communicated with the Divison, that did the 990 mini-computers? Did you share tools/processes/standards?

The 990 was not manufactured in Lubbock, but used widely across TI for automation control.
My first experience with the 990 computer was in 1978 working in IC parametric test as a repair person.  These controlled the testers for testing TI’s memory IC’s – TMS4044 and the TMS4116 (used in the 99/4A home computer). We reloaded programs using the paper IBM (Hollerith punch card) cards.
You could also use the 16 switches on the front panel to program in Hexadecimal code.  – Klaus – My brain is straining to remember back to Feb 1978 - lol

 

Q) The power transformer of the 99/4A was recalled. Did your group get blamed for that?

 

Klaus, it seems that I have the greatest memory of this in the 3 years working in Home Computers at TI.
This is the story as I remember it:
The power transformer was designed, manufactured by a 3rd party.
Used by other manufacturers of computers, games, etc....
There was an Incident of a power transformer on an Apple (Lisa) computer peripheral being shorted, which caused the transformer to over heat (melt) and start a house fire.
Since the bezel of the TI 99A was metal, shorting the power transformer was also reported to to cause a shock hazard risk. I personally tested 10 of these transformers – shorted the output and measured the power transformer’s plastic case temperature vs time.  NOT GOOD!!
TI started a recall / sending their 994/A customers a modified transformer.
If you have one, you will see the green / white “check” on the piece that has an inline fuse.  So, the “fix” was to place an inline fuse that will open if there is an over current (electrical short) situation -  Which I also tested to determine what the case temperature was when the fuse opened.
Apple was sued – TI took the corrective action and replaced over 400,000 power transformers with the attached inline fuse.   

Q) With the release of the 99/QI motherboards came an updated ROM v2.2 which prevented third-party modules (Atari, Funware,..) from running. What was your opinion to this step back in the days?

I was in involved in the QI (Quality Improved) release.  In fact, I was given one of these systems, with expansion box.  It had the beige colored top / bottom case and no metal bezel.  It looked “friendlier”.  My memory is that the motherboard was layed out neater (more efficient) and some circuits had been integrated, there fore less components.  TI obviously didn’t learn, they had similar software restrictions with their portable PC.  Was this so the owners would be forced to buy TI software?  I say “YES” and it was a bad marketing strategy.

 

Q) What do you know about the never released 99/4B and the 99/5?

 

No, Sir  Only the work being done in Lubbock on the 99/8.
These other products may have been developed in Dallas, where I am now.

 

Q) We lately encountered 99/4As manufactured for Argentina with a different TI title screen. Do you remember any exotica like that which you encountered during your work?

 

It seems like I remember seeing “Power up” screens in different languages.  This would have been normal for systems sold to non-English speaking countries.

 

Q) Did you test the Hex-Bus and its devices?

No, I did not – I think the CC40 was the first to incorporate TI’s hex bus

 

Q) Did you test the CC-40 Portable as well?

No, I remember hearing of this computer when I was leaving the Home Computer Division in Sept 1983 – LDC display, different processor, runs of AA batteries, etc  I might have seen one in later years in the office areas – I left TI in 1988

 

Q) At which stage of development was the 99/8, when you got involved in that project?
Q) Can you describe in detail your involvement with the 99/8?
Q) Do you have any nice stories about the 99/8?
Q) I was under believe that the 99/8 did not happen due to a political decision of TI to leave the home computer market. You wrote on the net that the units had RFI (Radio Frequency Interference) issues and that blocked the 99/8 from being manufactured. I do find that very interesting. Can you explain that in detail?
Q) Which part in the machine caused these? How long did these RFI block the 99/8 from production? Did you solve the issues before the internal Quality run was done?

 

Klaus, I will group these questions into one and this will test my memory –
As I remember, TI wanted to abandon the RF shield design of the 99/4A.
IF you have taken one apart, it is the metal shell that attaches to the mother board.  This caused a lot of problems on the 99/4A during drop tests, etc because the RF shield could move and short on other components and caused damage.

Of course, the RF shield is a Faraday shield which prevents the RF energy from the clock and video circuits from being transmitted into the air and interfering with other electronics which might be nearby.  It also acts to shield incoming RFI, but this is secondary.

On the 99/8, the initial design was to use a conductive spray paint that was heavy in Lead.  Initial testing showed very good results as a replacement for the bulky metal shield of the 99/4A.  However, this was about the time that the US EPA was determining that Lead was a health issue.  TI was not able to use the Lead coating that they had spent several months qualifying for RFI on the 99/8.
Going to an alternate solution would require time for testing and qualifying.
Atari, Commodore had released their products and TI was months away from a possible release.  This is what I remember from 32 years ago.
Does your 99/8 have a top and bottom case?  It would be interesting to know if it has RFI shielding and what method was used.

 

Q) What other hardware was developed or even tested, but never made it to the home computer market?

For home computers:  I was not involved in R&D.
For other products-  there was the CB radio, digital watches (manufactured in LED and LCD form for a few years), digital thermometers some stereo audio components.

 

Q) How did it came that you left TI in 1983?

 

I transferred to the semiconductor FAB at TI in Lubbock in Sept 1983 and I left TI in 1988 after 5 years in the FAB.  I moved to Austin, TX and worked for the R&D consortium – SEMATECH.  This was a great experience working with many bright minds of the semiconductor industry and meeting the co-inventor of the integrated circuit (CEO at the time) Dr. Robert Noyce.  Please read about Dr Noyce and Jack Kilby (co-inventor of the integrated circuit at TI).

Q) Are you still in touch with other TI employees from that time?

 

Yes, There are a few from Home Computers that I work with today.
I can ask Ron Powell about his memory of the 99/8.
He was involved with this product more than I.

 

Q) Have you ever visited any TI-99 User Group Meetings or been part of a Usergroup?

 

No, There was a User’s group and a magazine that came out of Round Rock, Texas in the early to mid 1980’s.  I subscribed to that.
I am sure this does not exist anymore.

 

Q) Have you kept any TI-99 related hardware, software or documents?

Good Question Klaus – I gave my 99/4A console, expansion box, speech synthesizer module, joysticks to a friend when I moved from Lubbock.  I had schematics for many years of the 99/A, but those are gone now.  I feel a little sad when I think of not keeping all of these items.

Q) Are you proud that these products still work after that many years?

Yes, These were expensive “toys” at the time.  I am happy that there are still
99/4A hobbyists keeping the spirit alive today.  I hope that there are spare power supplies (internal PCB) available today.  This was the #1 field failure of the 99/4A.
I think the quality was good – per quality standards for consumer electronics at that time.

Q) It seems you are employed with TI again since 2009. Can you describe what has changed since then? Are you in the same location as in the early 80's?

I will be 60 yrs old in Jan 2016 – I was 20 when I started working at TI.
TI was a household name – consumer electronics – calculators, watches, speak and spell.  TI had their profitable semiconductor division making memory and logic chips.  TI had a strong Defense electronics division. TI had a materials development division.  Many changes at TI Klaus – TI does their calculator manfucaturing in Asia, sold off all of their Defense work to Raytheon and other companies,  Closed all of their memory FABs in 1998.  It’s a different world and a different work force.  One thing remains the same – TI’s Quality in manufacturing is still world class.
I work at the Corporate headquarters in Dallas, Texas

 

Q) What are you doing now at TI?

 

I am an Engineer in one of TI’s Semiconductor FABs.

 

The 2nd set of questions (not displayed here) is yet to be answered.

[wolhess]

Hello Klaus,

I'm pretty impressed with what you have for contacts and what interviews you did.

Thank you for providing the interviews.

Very interesting!

[sometimes99er]

7 hours ago, kl99 said:

It was definitely not the Z80-based version, since that looked more like a high-end stereo component, including the wooden side panels. Compared to the 99/4 it was extremely fast, since the video was memory mapped (and you had a speedy processor).

Nice.

[Tursi]

So these are ongoing? Nice!

 

[GDMike]

Wonderful stuff!

[fabrizio]

19 ore fa, kl99 ha detto:

Herman Schuurman
di Klaus Lukaschek
Intervista realizzata nel dicembre 2015

 

Herman Schuurman ha lavorato 36 anni in Texas Instruments, dal novembre 1977 al suo pensionamento nel 2013. Nel marzo 1978 è stato promosso a capo programmatore per il gruppo di prodotti di consumo a Lubbock.

La descrizione del suo lavoro è tratta da LinkedIn per quella designazione:

Progettazione software per prodotti avanzati per personal computer. Progettazione e realizzazione del sistema di sintesi vocale basato sul sintetizzatore vocale TMS5200; Sistema di sviluppo mini memoria TI 99 / 4A; Sezione I / O del computer di casa 99/4; Sezione I / O dell'interprete BASIC; software di sistema per vari dispositivi periferici.

[ https://www.linkedin.com/in/herman-schuurman-60584b9/]

 

D) Com'è stato lavorare per TI nel Consumer Products Group?

 

E 'stato molto divertente. Lubbock è una comunità relativamente piccola (circa 180.000 quando vivevo lì), quindi abbiamo una squadra affiatata che si riuniva anche al di fuori del lavoro. Lo stesso Lubbock è simile al deserto: caldo secco d'estate e freddo d'inverno. Fa parte della pianura meridionale; piatto come un pancake, senza colline intorno di cui parlare. I nomi delle città circostanti riflettono questo: Levelland, Plainview, Shallowater, ecc ...

Personal Computer di Lubbock, ma ho lavorato per il Consumer Products Group di Almelo, Paesi Bassi, mentre il mio permesso di lavoro americano è arrivato attraverso il sistema. Nel marzo del 1978 mi sono finalmente trasferito a Lubbock, non avendo mai visto il posto.

 

D) Puoi descrivere la tua relazione con TI come dipendente? Ci hai quasi lavorato tutta la vita lì.

 

Fino a poco tempo fa, TI aveva molte diverse suddivisioni. Questo mi ha permesso di lavorare dai sistemi di consumo a quelli industriali alla ricerca, ecc. In definitiva TI è concentrato di più sull'analogico, quindi immagino che sarebbe più difficile rimanere in attesa per tutta la tua carriera e avere ancora una varietà di incarichi.

 

D) Come mai hai lasciato il Consumer Products Group alla TI nel 1981?

 

Sono entrato a far parte del gruppo con sede a Dallas da cui proviene originariamente Don Bynum, il Centro di ingegneria aziendale. Quando me ne sono andato, la maggior parte della progettazione del sistema era finita e il team stava passando alla progettazione delle applicazioni.

 

D) Puoi descrivere in dettaglio il tuo coinvolgimento con il sintetizzatore vocale TMS5200?

 

Il design originale per il TMS5100 è stato realizzato da Larry Brantingham, Paul Breedlove, Richard Wiggins e Gene Frantz. Gene era un capo del gruppo di discorsi (sede di Speak & Spell) quando sono unito a TI a Lubbock, e Larry è trasferito a TI France (Nizza) poco dopo. Alla fine mi sono ritrovato in un gruppo con Richard Wiggins quando sono entrati nel Corporate Engineering Center di Dallas.

La seconda generazione di TMS5100, TMS5200, è stata fornita per alimentare la periferica vocale. Il mio coinvolgimento con il sintetizzatore vocale è stato quello di creare il software di traduzione del testo e tesatura allophone in coordinamento con Kathy Goudie (che ha lavorato per Gene), che era responsabile della creazione del set allophone (suono) e delle regole di traduzione . L'articolo collegato di Sharon Crook è sostanzialmente una rivisitazione della documentazione interna sul software di sintesi vocale.

 

D) Tutti i sintetizzatori vocali TI-99 hanno la porta che era pensata per l'inserimento dei moduli Discorso, le prime unità hanno anche l'interfaccia per tali moduli. Tuttavia, non sono stati rilasciati moduli. Hai un'idea di questo? Puoi illuminarci con una storia su come questo è stato pianificato e abbandonato?

 

Il modulo vocale è stato fornito con 200 frasi bizzarre che potrebbero essere usate nei moduli software. C'è una disposizione per la ROM da aggiungere in seguito per espandere il vocabolario disponibile, ma l'aggiornamento della funzionalità di sintesi vocale ha reso questo punto controverso.

 

D) Puoi descrivere il tuo coinvolgimento nel progetto TI-99/4?

 

Il progetto Home Computer (99/4) è iniziato circa un anno prima del mio team di Lubbock. Credo che i promotori originali del progetto siano Granville Ott e Len Donohoe. Inizialmente ero stato assunto per lavorare sull'S-70, un piccolo computer scientifico, ma quando sono atterrato a Lubbock, quel progetto era stato spostato nel Data Systems Group di Austin, e sono stato messo a lavorare sull'SR-62, un piccolo computer autonomo che ha condiviso la maggior parte dei suoi software con il computer di casa. Oltre alle funzionalità Home Computer, l'SR-62 aveva un monitor piccolo incorporato e una stampante termica. Quando il computer di casa alla fine cadde in ritardo, il team SR-62 fu spostato per completare il 99/4.

Dato che il mio background era nella progettazione del sistema operativo, ho lavorato su molte cose relative agli I / O come l'audiocassetta, la stampante termica, ecc. Ero / a responsabile della sezione I / O dell'interprete BASIC, incluso l'I / O formattato O, ecc ... Bill Nale e abbiamo diviso quel progetto, con Bill responsabile dell'hardware e del software di livello basso, mentre io ho preso la progettazione e l'implementazione del file system. Questa è stata ricevuta volta in cui ricordo di aver avuto contatti con qualcuno di Microsoft, anche se molti siti Web 99/4 sembra pensare che Microsoft fosse responsabile di molti software sul 99/4.Bob Greenberg è uscito una volta per convalidare la progettazione del file system (non ci sono state modifiche alla progettazione).

 

D) La TI Dimensione 4 sembra quasi la TI-99/4 ed è del 1978/1979. Ne sai qualcosa a riguardo?

Link per Schuurman al thread atariage dimension4

 

Sembra proprio un 99/4 iniziale; Non ricordo il nome Dimensione 4, ma potrebbe essere un nome commerciale anticipato per il 99/4. Non era sicuramente la versione modificata su Z80, dal momento che sembrava più un componente stereo di fascia alta, compresi i pannelli laterali in legno. Rispetto al 99/4 è stato estremamente veloce, poiché il video è stato mappato in memoria (e hai avuto un processore veloce).

Prima di arrivare a Lubbock, c'era un po 'di lavoro su un chip GPL nativo, ma nel 1978 era stato sostituito con un design basato su TMS9985 a 8 bit. Sfortunatamente, quel chip non è mai funzionato correttamente, quindi alla fine abbiamo dovuto adattare un TMS9900 un 16 bit in un design un 8 bit. Se ti piace vedere maggiori informazioni sulla debacle del design 9985, controlla la pagina di Karl Guttag su http://www.kguttag.com/2013/08/10/if-you-havent-tested-it-it-doesnt-work /.

 

D) Hai ancora una TI-99? Se sì, puoi descrivere ciò che hai conservato. Quando è stata l'ultima volta che l'hai usato?

 

Qualche anno fa ho consegnato a Joe Zbiciak tutto il mio hardware e software 99 / 4a, dal momento che non lo toccavo da un po '. Joe è più interessato ai sistemi legacy, quindi ne ha fatto un uso migliore. Ho tenuto il monitor Panasonic, ma poche settimane fa è fallito, quindi ora sono totalmente fuori dai contenuti relativi a 99 / 4a.

 

D) Il sistema di sintesi vocale è stato riutilizzato per altri prodotti Discorso di TI o di altre società?

 

Non per mia conoscenza. Ricordo che Ute Marcotte stava lavorando a una versione tedesca del set di regole / allophone text-to-speech, ma non è così mai uscito. La maggior parte del successivo interesse ha riguardato il riconoscimento vocale indipendente dal relatore, su cui ho lavorato diversi anni dopo nel gruppo Telecom Systems.

 

D) Avere qualche relazione con Speech, qual è la tua opinione su Siri di Apple?

 

Rispetto ai primi lavori che abbiamo svolto sul riconoscimento vocale indipendente dal relatore, sia Siri che Cortana sono fenomenali. Naturalmente anche la quantità di potenza della CPU e lo spazio di memoria disponibile sono incomparabili. Uso un po 'il riconoscimento vocale su iPhone, ed è davvero buono.

 

D) Il tuo nome ha origine nei Paesi Bassi o in alcuni paesi vicini? Puoi illuminarci?

 

Sono nato in Olanda e trascorro i primi 24 anni della mia vita lì. Mi sono laureato alla Technische Hogeschool Twente (ora nota come Università di Twente) a Enschede, Paesi Bassi. Dopo la laurea sono iscritto a TI e sono trasferito negli Stati Uniti. Secondo la mia laureato in Ingegneria elettrica, gran parte del mio background era in sistemi integrati e progettazione / implementazione di sistemi operativi, motivo per cui sono stati assunti da TI.

 

D) Sai come è arrivato il comando per caricare un programma si chiama "OLD" sulla TI?

 

Probabilmente provata dall'originale Dartmouth BASIC, dove VECCHIO usato per recuperare un programma dalla memoria e NUOVO per combinato un nuovo programma.

 

D) Bill Cosby era il volto principale della TI-99 per il pubblico, era anche una persona a cui i dipendenti guardavano?

 

La pubblicità è stata gestita dal team di marketing, ma Bill Cosby era molto popolare a quei tempi, sia dalla sua precedente serie I Spy, sia attraverso lo spettacolo Fat Albert e Cosby Kids. Inoltre, il dottorato in Scienze dell'educazione non ha fatto maschio.

 

D) Qual è lo stato del ruolo di Don Bynum con cui hai lavorato?

 

Don stava lavorando a una riprogettazione della 99/4 mentre era nel Corporate Engineering Center di Ti Dallas (il Ranger). Alla fine del 1980 si trasferì un Lubbock per rilevare il gruppo di computer di casa da Pete Bonfield (che divenne presidente e amministratore delegato di ICL in Inghilterra). Ha guidato la 99 / 4a e gli atti della scatola periferica, e ha effettuato la 99/2 e 99/8. In realtà non ho mai visto, ma si dice che avesse un vecchio cellulare per pianoforte con la catena periferica 99 / 4a che occupava collegato mobile.

Pochi mesi dopo l'acquisto della 99 / 4a mi sono trasferito dal team Lubbock al vecchio gruppo di Bynum, il Corporate Engineering Center di Dallas.

 

D) Era chiaro fin dall'inizio che ci sarebbe stato un software di sintesi vocale o un'idea brillante da parte di un dipendente?

 

Non so chi abbia avuto l'idea del text-to-speech con il gruppo vocale, e neanche Kathy ricorda. Controllerò con Gene Frantz per vedere se ricorda ancora. Tieni presente che non era scontato che il text-to-speech avrebbe effettivamente funzionato in modo accettabile nel 9900 fino a quando non lo avessimo effettivamente implementato.

 

D) Quindi potremmo dire che il tuo lavoro sul Text to Speech ha reso irrilevanti i moduli vocali per il sintetizzatore vocale, giusto?

 

Sì.

 

D) Su quale lavoro scientifico è stato basato il tuo software di sintesi vocale? C'è qualche linguista o scienziato fonetico coinvolto?

 

TI aveva già un team vocale attivo situato nel Lubbock (uno dei loro prodotti più noti è probabilmente Speak N Spell). La persona con cui ho lavorato, Kathy Goudie, ha un dottorato di ricerca. in Linguistica e gli allophone utilizzati nel software di sintesi vocale sono stati creati da un ampio database vocale che era già stato sviluppato all'interno del team vocale.

 

D) Non sono riusciti a trovare molte informazioni sui computer SR-62 e SR-70, sono mai stati rilasciati? Quali sono state le specifiche dell'SR-62? Il suo software con il 99/4, questo potrebbe essere interessante per la nostra gente.

 

Non sorprende, dal momento che si tratta di designatori interni e nessuno dei due prodotti è arrivato sul mercato. L'era SR-62 era essenzialmente un computer domestico autonomo con un monitor integrato e una stampante termica.

 

D) Sai quanto Microsoft o Bill Gates sono stati coinvolti nella TI Basic / System Rom del 99/4?

 

Microsoft non è stata coinvolta nello sviluppo del 99/4. Loro (nella forma di Bob Greenberg) furono incaricati di sviluppare BASIC per SR-70 (che a volte viene anche chiamato 99/7), ma BASIC per 99/4 fu sviluppato internamente.

 

D) Perché il chip GPL nativo è stato sostituito con il TMS9985? Fino a che punto è stato sviluppato il chip GPL?

 

Sebbene fosse prima del mio arrivo a Lubbock, il chip GPL doveva essere sviluppato per un cliente esterno. Quando quel cliente ha abbandonato, anche il chip GPL è stato abbandonato e sostituito dal TMS9985.

 

D) L'accesso ai dispositivi periferici tramite un'interfaccia DSR ci consente di collegare dispositivi moderni alla TI senza modificare la ROM di sistema. Sai chi ha avuto l'idea su questo? TI lo aveva già prima del 99/4?

 

È passato troppo tempo per ricordare i dettagli esatti, ma probabilmente è stato risolto da Bill Nale e me. Bill avrebbe gestito le funzionalità di livello 1 (comunicazioni hardware, lettura / scrittura settoriale, formattazione del disco) e avrei gestito le funzionalità relative ai file. Comprendere l'hardware è stato sviluppato contemporaneamente al software, mi ha permesso di lavorare sul codice del file system simulando la routine di basso livello su un minicomputer TI-990, usando un file di grandi dimensioni sul disco rigido del minicomputer per simulare un disco floppy .

Tieni presente che ho già sviluppato un sistema simile per un sistema Intellec-80 (basato su Intel 8080) usato floppy da 8 "come parte della mia tesi.

 

D) Come ti sei sentito quando hai saputo che TI uscirà dal mercato dei computer domestici dopo essere stato coinvolto nella configurazione?

 

Triste, anche se probabilmente era inevitabile a causa delle perdite finanziarie. Sono comunque contento di aver avuto la possibilità di lavorarci su.

 

D) Qual era la politica di TI con hardware / software / documentazione / schemi su progetti annullati? Sembra che alcuni fortunati dipendenti abbiano ottenuto una TI-99/8 quando TI ha lasciato il mercato.

 

In genere devi ottenere l'autorizzazione ufficiale per ottenere legalmente qualsiasi di queste cose. Detto questo, è tutto il possibile che la gestione in quel momento abbia regalato alcuni cimeli.

 

D) Come ti senti quando senti che quelle macchine sono ancora in funzione dopo tutti quegli anni e ci sono persone che fanno ancora cose con loro?

 

Assolutamente stupito e felice. Non avrei mai immaginato che il 99/4 sarebbe sopravvissuto per oltre 35 anni, soprattutto con il tipico cambio rapido nell'era dei computer.

 

I migliori saluti,

Herman Schuurman

Great Klaus, thank you for sharing these interviews.

[sometimes99er]

19 hours ago, kl99 said:

Before I arrived in Lubbock, there was some work done on a native GPL chip, but by 1978 that had been replaced with an 8-bit TMS9985 based design.  Unfortunately, that chip never ran correctly, so we had to eventually fit a 16-bit TMS9900 into an 8-bit design.

19 hours ago, kl99 said:

It was definitely not the Z80-based version, since that looked more like a high-end stereo component, including the wooden side panels.  Compared to the 99/4 it was extremely fast, since the video was memory mapped (and you had a speedy processor).

19 hours ago, kl99 said:

Q) With the release of the 99/QI motherboards came an updated ROM v2.2 which prevented third-party modules (Atari, Funware,..) from running. What was your opinion to this step back in the days?

I was in involved in the QI (Quality Improved) release.  TI obviously didn’t learn, they had similar software restrictions with their portable PC. Was this so the owners would be forced to buy TI software?  I say “YES” and it was a bad marketing strategy.

So the GPL processor didn't make it. The 9985 didn't make it in time. The Z80 works like a charm, but someone doesn't like it, and they quite successfully "erase it". They have to shoehorn the 9900 in. Basic runs slow. You couldn't run Assembly on the stock machine - unlike all the other competitive computers. They're making it relatively difficult for third party developers. GROM and GPL also secured a relatively closed architecture. Their expansions are relatively expensive. Do I have to mention the joystick port. And at some point on the curve of success (units selling), they sell the 99 at a loss, and they're not making up for it.
- So with all this "TI only and keeping it to yourself", I think they could certainly have made a general consensus for "erasing" the Microsoft involvement. Whatever, Microsoft probably had the longest laugh. - Now, we could try and blame Microsoft for TI Basic running so slow.

[GDMike]

I suppose they really didn't want to interfere with their other professional line at the same time.

[+hloberg]

removed

Edited September 17, 2019 by hloberg

[+mizapf]

By the way, I am glad that there never was a GPL-based CPU. GPL would have been an extremely complex instruction set architecture with forbiddingly long microprograms. It may look interesting as the virtual architecture that it actually became, but in silicon?

[Tursi]

5 hours ago, mizapf said:

By the way, I am glad that there never was a GPL-based CPU. GPL would have been an extremely complex instruction set architecture with forbiddingly long microprograms. It may look interesting as the virtual architecture that it actually became, but in silicon?

Agree with that, at least if it was supposed to parse the byte code as it stands today, or even a subset of it.

[Elia Spallanzani fdt]

in this post ( http://www.mainbyte.com/ti99/history/history.html) the author claims that the sales of the ti99 were good when suddenly "the disaster struck":

"It was discovered that under certain circumstances, a person might receive an electrical shock from the computer’s power supply, although no user had yet experienced such a problem. "

 

doesn't it seem like a sort of self-absolution for a series of technical and strategical errors?

Edited November 16, 2019 by Elia Spallanzani fdt

[RXB]

On 9/17/2019 at 8:38 PM, Tursi said:

Agree with that, at least if it was supposed to parse the byte code as it stands today, or even a subset of it.

Well this comes down to a 16 bit CPU being hamstrung with nothing but 8 bit chips all around it to do anything.

Nothing like tying one hand behind back and attempting to work as hard as any two handed person.

 

GROM/GRAM on a 16 bit bus would have been pretty hot to watch work, instead we got the nightmare version.

This resulted in our TI Basic, and all the GROM carts.

[Tursi]

On 11/16/2019 at 7:50 AM, RXB said:

Well this comes down to a 16 bit CPU being hamstrung with nothing but 8 bit chips all around it to do anything.

Nothing like tying one hand behind back and attempting to work as hard as any two handed person.

 

GROM/GRAM on a 16 bit bus would have been pretty hot to watch work, instead we got the nightmare version.

This resulted in our TI Basic, and all the GROM carts.

I still think it wouldn't make much difference. I tested the performance of the interpreter vs the actual GROM access (remember that the interpreter runs out of 16-bit ROM so is already running as fast as is possible, and only periodically accesses GROM to get more instructions). The tests back in 2015 suggested that GROM access was only about 8% of the performance. So even if you got GROMs working with zero-wait-state 16-bit access, without an interpreter re-write you'd only see 8% performance increase.

 

https://atariage.com/forums/topic/237554-internal-gpl-interpreter/?do=findComment&comment=3228065

I'm still pretty sure we could do far better just re-writing the interpreter with modern knowledge. I'm just not ready to prove it.

 

[Omega-TI]

3 minutes ago, Tursi said:

So even if you got GROMs working with zero-wait-state 16-bit access, without an interpreter re-write you'd only see 8% performance increase.

This hobby is something different to different people (of course), but for me personally, even if there was a way to hot rod the motherboard & CPU for faster processing I wouldn't do it.  Don't get me wrong, I LOVE seeing all the new gadgets and all the ways people have found to "work around" the TI's limitations and hanging new goodies off the console is part of the fun, but for me personally, the CPU and anything related to it is sancrocanct.

 

I'll use the Star Trek example of Veger, the thing was awesome and immense, but at the core was a piece of old obsolete tech!  Take that away, and it does not feel the same.

[apersson850]

Since I was interested in hardware too, when I was using my TI 99/4A a lot, I implemented the all 16-bit RAM in my console. Pascal runs about 10% faster when all RAM is 16-bit wide in the machine. The main increase probably comes from things like that parts of the p-system code is loaded into 8 K RAM expansion. The 80-column screen memory is also there. Thus, some p-system runtime is spent in this memory. The inner parts of the PME are transferred to the internal RAM PAD, so they are 16-bit already. The major part of the interpreter is on the p-code card, which is 8-bit access reagardless of the RAM. User's code is running from VDP RAM, space permitting.

[RXB]

16 hours ago, Tursi said:

I still think it wouldn't make much difference. I tested the performance of the interpreter vs the actual GROM access (remember that the interpreter runs out of 16-bit ROM so is already running as fast as is possible, and only periodically accesses GROM to get more instructions). The tests back in 2015 suggested that GROM access was only about 8% of the performance. So even if you got GROMs working with zero-wait-state 16-bit access, without an interpreter re-write you'd only see 8% performance increase.

 

https://atariage.com/forums/topic/237554-internal-gpl-interpreter/?do=findComment&comment=3228065

I'm still pretty sure we could do far better just re-writing the interpreter with modern knowledge. I'm just not ready to prove it.

 

Well the reason for GPL in the first place was the sheer amount of RAM it saved vs loading everything which is the number one plague of the PC, lack of memory.

I guess you could also claim STANDARDIZATION also as who needs 40 ways of doing same thing and not many are compatible. 

[+FarmerPotato]

6 minutes ago, RXB said:

Well the reason for GPL in the first place was the sheer amount of RAM it saved vs loading everything which is the number one plague of the PC, lack of memory.

I guess you could also claim STANDARDIZATION also as who needs 40 ways of doing same thing and not many are compatible. 

The reason for GROMs is that they were 16 pin packages, vs 28 for a ROM. In packaging, this was a big cost savings. In late 1982, a 4K ROM in the 4A still had an internal cost of $4, whereas GROM was $1.33.  My source on this is Ron Wilcox, engineer of the QI console. 

 

TI pioneered the use of latching an address as 2 separate bytes to save half the pins (this might have even been one of Karl Guttag's patents in the 70s? or maybe I'm thinking of dual-port VRAM.) The 4116 DRAM did this (albeit with no auto-increment). And "two address writes, half the pins" became the industry standard as memories got larger, and hungrier for more address pins. 

 

Also, I'm not sure what process was used to create the ROMs, but I know that the GROM chips manufacture required only one customer mask layer on top to program the bits.

 

[urbite]

On 11/18/2019 at 9:08 AM, FarmerPotato said:

The reason for GROMs is that they were 16 pin packages, vs 28 for a ROM. In packaging, this was a big cost savings. In late 1982, a 4K ROM in the 4A still had an internal cost of $4, whereas GROM was $1.33.  My source on this is Ron Wilcox, engineer of the QI console. 

 

TI pioneered the use of latching an address as 2 separate bytes to save half the pins (this might have even been one of Karl Guttag's patents in the 70s? or maybe I'm thinking of dual-port VRAM.) The 4116 DRAM did this (albeit with no auto-increment). And "two address writes, half the pins" became the industry standard as memories got larger, and hungrier for more address pins. 

 

Also, I'm not sure what process was used to create the ROMs, but I know that the GROM chips manufacture required only one customer mask layer on top to program the bits.

 

It was conveyed to me that the GROMs were (relatively) inexpensive PMOS chips with a calculator heritage, and were used because of the low cost.

[sometimes99er]

https://atariage.com/forums/topic/164068-was-gpl-a-mistake/?do=findComment&comment=2098697

[Davvel]

Hi I have a curious question for the GPL experts. What did the original TI developers use to code in GPL? I mean did they have GRAM chips hooked up to their TI 99 4A and after they finish all the testing they burn their byte code on GROMS using some sort of eprom programmer which was linked to this whole super system? I am interested in those who worked before GRAM Kracker was invented,  I mean 1980 and 1981. 

[+Ksarul]

TI did their development on a TI-990 computer with a GPL Assembler and a standard Assembler for almost all of their coding (a few Assembly-only packages were actually programmed on the TI with an Editor Assembler). Software could be tested several ways: there were EGROM cartridges that had space for GRAM simulation circuitry and EPROMs to contain the data, and EGROM box that sat beside the console and plugged into the cartridge slot (this method could simulate five GROMs as opposed tot he one that could be emulate on an EGROM cartridge), a GSIM box that emulated GRAM chips and which could be loaded with programs like a GRAM Kracker (a big silver pizza-box that plugged into the cartridge port using a cable), and a PEB GRAM card which plugged into the cartridge port too. All of these items were from TI. Later, there was the Mechatronic GRAM Karte, the p-GRAM card, the GRAM Kracker, the GRAMULATOR, the Maxi-Mem, the Wiesbaden Supermodul (versions I and II), the Karlsruhe Module, and the Baunatal Module. 

[Davvel]

On 12/20/2019 at 1:59 AM, Ksarul said:

TI did their development on a TI-990 computer with a GPL Assembler and a standard Assembler for almost all of their coding (a few Assembly-only packages were actually programmed on the TI with an Editor Assembler). Software could be tested several ways: there were EGROM cartridges that had space for GRAM simulation circuitry and EPROMs to contain the data, and EGROM box that sat beside the console and plugged into the cartridge slot (this method could simulate five GROMs as opposed tot he one that could be emulate on an EGROM cartridge), a GSIM box that emulated GRAM chips and which could be loaded with programs like a GRAM Kracker (a big silver pizza-box that plugged into the cartridge port using a cable), and a PEB GRAM card which plugged into the cartridge port too. All of these items were from TI. Later, there was the Mechatronic GRAM Karte, the p-GRAM card, the GRAM Kracker, the GRAMULATOR, the Maxi-Mem, the Wiesbaden Supermodul (versions I and II), the Karlsruhe Module, and the Baunatal Module. 

Thanks Ksarul for your detailed explanation. I think you summed it up really well. In fact I am amazed in how much detail you could remember, impressive.  

[+Ksarul]

51 minutes ago, Davvel said:

Thanks Ksarul for your detailed explanation. I think you summed it up really well. In fact I am amazed in how much detail you could remember, impressive.  

It actually helps that I have most of these hardware items in my collection.