Halt and Catch Fire

[calexanian]

Just watched the premier. Ok, so they are taking one little semi fact and expanding it into a fictional story. Ok... Practically no bearing on reality. That being said it looked good. We shall see how close they get to the real story of how CDP engineered the bios. I have a feeling it will be nowhere close and purely done for entertainment value. Also people who know IP know that If IBM had actually contacted CDP and barged into their office like that it would have been a non starter for a lawsuit and any competent lawyer would have immediately sued on the claim of declaratory judgment.   Anyways, I will be watching

[T3sl4co1l]

I found it funny the way they made a big deal out of having to manually write out all 65536 bytes of the IBM BIOS. That's the easy, if laborious bit. The crucial and very labour intensive part is decoding all that data into assembly language and then figuring out what it does so that a specification can be written. That would have taken much longer and was the crucial step, and would have made more interesting TV too I think.

If you're the kind of person who likes CSPAN maybe

Tim

[jlmoon]

This Subject had me doing a double take.  I thought someone might have been reflecting on the old Moto 6800 engineering address line test mode..(the infamous undocumented Op Codes from the past)  Many a day spent  my head with these issues. 

JLM

[echen1024]

I thought it was pretty good, but as mojo-chan said, the 65536 numbers was not all that difficult. It's just quite repetitive. Looking forward to ep. 2

[jlmoon]

I have never been able to find any evidence that this was a true HCF mode though. By "true" I mean it physically broke something, not necessarily started a fire. There were rumours that it could make the address line traces burn, but I can't really see how. It's not like the old PET monitor that could actually be killed by certain instructions.

No, I suspect (don't quote me on this) it was utilized as a address bus debugger (by engineering - hince the low level of documentation on the OpCode) or sequencer to test for errant designs in interaces.  Maybe it was some razzle dazzle to blink all the address lights on the Altair 6800 front panel, don't know!.  I can't recall all the specifics on the actual issues when the 6800 wound up in this strange mode, but do remember it was a head scratcher when looking for the coding problems.

[N2IXK]

I had to roll my eyes all through that "reverse engineering" scene. Some of the issues I caught:

Cutting a beer can apart with a soldering iron (a Radio Shack cheapie at that).

"Checking Logic Levels" with a 60s vintage Heathkit scope that just showed 60 Hz hum jumping around.

Desoldering the BIOS chip to get the data off of it (and then showing the guy pulling the chip from the DIP socket it was actually installed in).

All those LED colors that didn't exist in 1983.

The flashes and pops when something "shorted out" looked more appropriate to a tube amplifier than a 5Volt PC motherboard.

Manually reading out the data from every ROM address location in HEX from the blinkenlights, reading it out loud, than having somebody else write it down, then reentering it all into the TRS-80--all without any transcription errors!  Guess doing a dump from an EPROM programmer wasn't sexy enough?

And what was the whole point of the exercise, anyway?  IBM published the BIOS code in the 5150 technical reference manual for anyone to see.

I have seen a LOT of reviews that are praising this show for its "technical accuracy".   

I have to wonder if doctors have similar WTF moments watching medical shows on TV.  Presumably they screw up the technical details there just as badly...

[Thilo78]

Just watched the premier.

I quit watching half-way through.
Maybe I missed some good part, but I was so bored with the pilot... 

I don't think that I'll be picking up later.
What's the actual point of the story? 

[mtdoc]

I have to wonder if doctors have similar WTF moments watching medical shows on TV.  Presumably they screw up the technical details there just as badly...

Yep, but it's usually even worse.

[W7NGA]

It's been a while ... but I bet those guys kicked themselves when they found out DOS 1.0 had a DEBUG hex dump utility.
Loved the oscilloscope .. blew my coffee out my nose when I saw that! 

W7NGA

[calexanian]

CDP did had all of the information since IBM made it available for other companies to product peripherals much the same as other companies had. They were just the first to run the copyright self developed gambit and won. It was a bunch of not very sexy engineers working round the clock to see how programs were supposed to run, and figure out how to get the same result without looking at the original. Laborious, but not impossible. Would have not made good television.

[Stonent]

I remember seeing red, green and yellow LEDs, all of which I remember getting in the radio shack "assorted LEDs" pack in the 1980s.

[calexanian]

The first LED's I remember were amber.  They had amber, green and red in the 70's

[SeanB]

And in the 1970's the best LED was a brightness that would be the equal of a wheat grain filament lamp operated at half voltage, and if you went over 20mA they would cremate themselves. In a TV studio lighting situation you would never be able to tell if they were on or off.

[kfitch42]

I found it funny the way they made a big deal out of having to manually write out all 65536 bytes of the IBM BIOS. That's the easy, if laborious bit. The crucial and very labour intensive part is decoding all that data into assembly language and then figuring out what it does so that a specification can be written. That would have taken much longer and was the crucial step, and would have made more interesting TV too I think.

From http://en.wikipedia.org/wiki/Columbia_Data_Products :

CDP introduced the MPC 1600 "Multi Personal Computer" in June 1982. It was an exact functional copy of the IBM PC model 5150 except for the BIOS which was clean roomed. IBM had published the bus and BIOS specifications, wrongly assuming[citation needed] that this would be enough to encourage the add-on market but not enough to facilitate unlicensed copying of the design.

Assuming Wikipedia was right, and the BIOS was "clean roomed," then I would expect CDP to actively avoid dumping the BIOS chip (manually or otherwise) to avoid even the appearance of copyright infringement.

[calexanian]

They wrote their own bios basses on the specifications. They had to find engineers that had never worked on such a project with no prior knowledge. Lawyers were involved every step of the way to insure it would hold up in court.

[retrolefty]

I found it funny the way they made a big deal out of having to manually write out all 65536 bytes of the IBM BIOS. That's the easy, if laborious bit. The crucial and very labour intensive part is decoding all that data into assembly language and then figuring out what it does so that a specification can be written. That would have taken much longer and was the crucial step, and would have made more interesting TV too I think.

If you're the kind of person who likes CSPAN maybe

Tim

Even less labor then that in that the original BIOS was contained in a 8KB ROM chip:

U33 BIOS = 8K

[baljemmett]

Assuming Wikipedia was right, and the BIOS was "clean roomed," then I would expect CDP to actively avoid dumping the BIOS chip (manually or otherwise) to avoid even the appearance of copyright infringement.

That's now how it works. One guy dumped and decoded the BIOS back into assembler, figured out what it did and wrote a specification for it. That spec was then given to someone else who had never seen the disassembly to re-implement from scratch.

The tale I've always heard was that the person/group writing the specification did so from IBM's published, commented BIOS listing (it's an appendix in the Technical Reference Manual), rather than by dumping and disassembling an actual ROM.  Then, as you say, it got passed to the 'clean room' where someone else re-implemented it from the spec.

[cloudscapes]

How accessible were EPROM programmers back then (I donm't know, I'm a young-un).

Testing logic high/low with a scope was pretty funny/embarrassing.

I remember seeing red, green and yellow LEDs, all of which I remember getting in the radio shack "assorted LEDs" pack in the 1980s.

Some orange ones too if I remember correctly. Plus there were those deep red ones that were almost maroon.

[JoeO]

This Subject had me doing a double take.  I thought someone might have been reflecting on the old Moto 6800 engineering address line test mode..(the infamous undocumented Op Codes from the past)  Many a day spent  my head with these issues. 

I have never been able to find any evidence that this was a true HCF mode though. By "true" I mean it physically broke something, not necessarily started a fire. There were rumours that it could make the address line traces burn, but I can't really see how. It's not like the old PET monitor that could actually be killed by certain instructions.

The original IBM 5150 had a poorly designed monochrome monitor.  In fact, if the 5150 BIOS initialization code hung in a certain section, the monitor was left in a state that would permanently take out the sweep transistor.  No sparks though.

[macboy]

...
I have to wonder if doctors have similar WTF moments watching medical shows on TV.  Presumably they screw up the technical details there just as badly...

I imagine they do. Do you think that a lawyer could watch "Suits" without rolling his eyes or outright cursing at the screen? I don't think so. I roll my eyes at that show every episode and I am an EE, not a lawyer.

[TerraHertz]

There doesn't seem to be a copy of the 1st episode on youtube yet. Just commentaries on it. So I can't see it. What's the theme?

CDP did had all of the information since IBM made it available for other companies to product peripherals much the same as other companies had. They were just the first to run the copyright self developed gambit and won. It was a bunch of not very sexy engineers working round the clock to see how programs were supposed to run, and figure out how to get the same result without looking at the original. Laborious, but not impossible. Would have not made good television.

If the theme is 'reversing the IBM PC, in 'clean-room' style to avoid copyright lawsuits' then the series is rubbish right from the start. There were so may copies of the early IBM PC, and all the ones I looked at had exact copies of the BIOS, even including the '(c) IBM' strings in the binary.
So far as I recall, IBM never even suggested the idea of suing anyone for infringement.

Furthermore, in those days IBM had PCB and chip manufacturing technology far ahead of everything else. They had dense multilayer PCB capability, and die-mount packages in which the chips were solder-bump flip mounted. If IBM had wanted to make the early PCs extremely difficult to reverse engineer and replicate, they easily could have. Instead they used bog-standard production technology, published the full schematics (I still have my complete service manual for the XT), and also made most of the BIOS source code available.

Even back then, IBM's total absence of engineering and legal effort to protect their Intellectual Property in the PC was so remarkable, that it was hard to see how there could be any explanation other than that IBM wanted everyone to copy their IBM PC architecture.

Which is an interesting idea, with even more interesting ramifications.

Incidentally the first LEDs were red and only red. None too bright either. It was several years before any other colors appeared.  Early LEDs tended to look like these. Yes, that's a formed metal rim around the base, and painted dot to mark the cathode.

[Stonent]

This talks about Compaq and CDP.

http://www.theverge.com/2014/6/9/5792656/close-up-halt-and-catch-fire-season-1-episode-2

When I first heard about the show, I thought it was supposed to be about Compaq, since it was based in Texas like Compaq was.  I find it strange that they make it like it's such a remarkable thing that they were going to clone the PC if it had already been done.

Also the scene where he said "Let's put a handle on it and make it portable."

[Stonent]

It looks like there are torrents for ep 1 and 2.

[calexanian]

Looks like its going to become more of a drama about now messed up the main character is.

[oPossum]

This talks about Compaq and CDP.

http://www.theverge.com/2014/6/9/5792656/close-up-halt-and-catch-fire-season-1-episode-2

When I first heard about the show, I thought it was supposed to be about Compaq, since it was based in Texas like Compaq was.  I find it strange that they make it like it's such a remarkable thing that they were going to clone the PC if it had already been done.

Also the scene where he said "Let's put a handle on it and make it portable."

It's fictional and not about any specific company, but the talk about the '186, making it faster than the PC, and a large Texas company made me think it may be based on the development of the Tandy 2000.