Prevue Guide Forums

First of all, very interesting work, both of you. Really neat to think these programs might one day be fully operable in simulators by fans of famous classic software, thanks to your combined efforts. (Big tip of the tophat to tin for his ubercracking skills especially!)

Anyway, just a couple comments:

Ari -

It's possible that the Amigas also have some sort of hardware demodulator in some sort of card that decodes the Prevue feed into the same format, but it's also possible that the demodulation is done in hardware, in which case it would work a little differently.

That's right, the A2000s used hardware demodulators that were manufactured in the form of ISA cards. The A2000 motherboards' ISA slots by Commodore's design weren't active (no data bus connectivity) and only provided power to whatever ISA cards were plugged into them. So UV just took advantage of that by building demodulators for the A2000s in the form of ISA cards, so the A2000s wouldn't need outboard ones. If you want additional details, I posted about this while still editing Wikipedia after someone nudged me about getting a couple details wrong re: A1000 and Atari hardware:
http://en.wikipedia.org/wiki/Talk:TV_Gu ... _equipment.

The strings present here must be compressed or obfuscated inside the binary

Have you considered that it's not the strings specifically that are compressed or obfuscated, but that the entire binary might simply be packed with some Amiga equivalent of PKLITE or UPX? Given the 2400 baud payload delivery method, executable binary compression would have made a lot of sense to UV. I doubt they would've developed their own binary packer, either. Surely they just used some off-the-shelf Amiga equivalent of PKLITE. That said, for all the popular binary packers (and even for many of the obscure ones), you can almost always find utils to unpack them (restore them to still-executable, uncompressed form). On that note, you might ask around in the Amiga world if there's any Amiga equivalent to Ben Castricum's UNP (a universal unpacker that understands many PC DOS executable packing formats). Or ask them to look at the PG executable files directly, assuming any of them would know how to identify and undo whatever packing method is being used on them.

Tin -

I think Atari EPG Jrs were still in operation a long time after the introduction of Amiga Prevue Guide, and it would make sense for the both to use the same data.

FYI: Prevue Guide's analog C-band feed was shut down and replaced with a digital (MPEG) feed when the yellow grid/WinNT stuff replaced the Amigas. Since there's no analog VBI in digital video, that left only WGN's analog satellite feed to feed the EPG Jrs' (and any surviving EPG Srs') demodulators with.

What effectively killed off the EPG Jr. units was WGN itself deciding to no longer offer an analog satellite feed about 5 years ago. When that happened, the company had three choices: find some other cable network that was still analog (and planned to stay that way for a long time), and pay its asking price for carriage ($$$$$$); build new add-on hardware for an entirely new (non-TV-based) method of delivering data to the Jr's ($$$$$$); or decommission the Jr's. Considering how few people were still using the Jr's by the mid-2000s, the more economical decision was the last one.

Can't remember where I downloaded it from, but I'm attaching a decommission report that came with someone's Jr. Shows several of its past users. Looks like the last two were SMATV providers (i.e. penny ante private cable service crap).

I am however sure that the Prevue Guide version will look for and need to see the genlock and audio switcher hardware at least, as the code will definitely control them both. I'll have to find one that I can buy and get imported to the UK!!

Correct. To run Prevue Guide properly, a real A2000 with authentic genlock and switcher cards would be needed. Just so the software would see them and not choke (at least), and (at most) so you could feed them real video and audio.

The only alternative would be to patch the Prevue Guide software so it made no calls/inqueries to the genlock and switcher cards -- and (just for cosmetics) possibly further patch it to just show graphical/text ads in the top half all the time.

Obviously, no actual hardware is needed to run the EPG Sr. or Jr. softwares. They just need to be fed the expected data via serial. If you have two PCs each with real serial ports, you could feed data to them with a serial cable that way as well (one PC acting as source etc.). Or, one PC, with a PCI card offering COM1 and COM2 ports.

Oh. P.S., for anyone interested: http://www.megaupload.com/?d=BJ1982SH

@swest77 - very good to hear from you I knew AriX was in touch with you and read both the info you put on Wikipedia and the talk page. I want to say thanks for providing the world with information about a system that so many people must have seen, but so few seem to know anything about.

Very interesting to read about the jr lasting for so long! Such a shame the general closedown of analogue removed the possibility of it continuing. Given that the signal was shut down making them usless, they were installed in small SMATV providers, and that it was so relatively recent, is there the possibility there's a number of these still installed and running in places, even though the output is useless, or did UV have a very strict decomissioning procedure?

Also the decomissioning report sheds some light on what I now know is called a select code (the NJJR6 bit that I so far thought was some kind of serial number)

If/when I get to the end of the atari disassembly I will try the amiga one, but I guess the coding will be an order of magnitude harder to reverse-engineer, as we think it was written in C. For now though AriX seems to be turning up all kinds of interesting information, so perhaps more light will be shed before such reverse engineering becomes necessary!!

Unfortunately the megaupload link says it's unavailable ATM, I will try and download ASAP.

Atari reverse-engineer is coming on apace, I'll update that thread now

swest77 wrote:

The strings present here must be compressed or obfuscated inside the binary

Have you considered that it's not the strings specifically that are compressed or obfuscated, but that the entire binary might simply be packed with some Amiga equivalent of PKLITE or UPX? Given the 2400 baud payload delivery method, executable binary compression would have made a lot of sense to UV. I doubt they would've developed their own binary packer, either. Surely they just used some off-the-shelf Amiga equivalent of PKLITE. That said, for all the popular binary packers (and even for many of the obscure ones), you can almost always find utils to unpack them (restore them to still-executable, uncompressed form). On that note, you might ask around in the Amiga world if there's any Amiga equivalent to Ben Castricum's UNP (a universal unpacker that understands many PC DOS executable packing formats). Or ask them to look at the PG executable files directly, assuming any of them would know how to identify and undo whatever packing method is being used on them.

Thanks for joining the forum!

It's entirely possible (in fact, likely) that the binary was compressed with UPX or something. The only reason I hadn't tried extracting it was because it wasn't compressed with PowerPacker 2.0, which is what nearly every other file on the system was compressed with. I will take a look into it tomorrow and see what I can find

@tin - My pleasure, re: the Wikipedia. I became interested in the EPG Sr. software (followed by the Prevue Guide software) as a kid because it's what made me aware that there were much better computers out there than IBM PCs and Apple ][s. "Wow, this thing has a GUI OS and high color graphics and it can do multi-tasking and smooth-scrolling text and does video and audio and..." Pretty fascinating thing to stare at when you're trapped in a world of 80x25 text mode and crappy Apple II graphics. I ended up having an Amiga 500 as a teenager and it was very strange in deed that I ultimately ended up working (however briefly and indirectly) for the company whose little programs turned me on to better hardware.

Anyway, guess I'll have to update the Wikipedia article again as you guys uncover more and more stuff. I already know now that I was mistaken in writing that the EPG Sr.'s top-screen banner (e.g. "BOB'S CABLEVISION") was configured via the local operator ESC menus.

Very interesting to read about the jr lasting for so long! Such a shame the general closedown of analogue removed the possibility of it continuing. Given that the signal was shut down making them usless, they were installed in small SMATV providers, and that it was so relatively recent, is there the possibility there's a number of these still installed and running in places, even though the output is useless, or did UV have a very strict decomissioning procedure?

I would assume UV ordered them all returned. Big companies can be very paranoid about proprietary systems staying adrift (especially when there's the intellectual property on them), even when it's outdated and almost nobody could want it or gain anything by having it. The odd thing is, they're greedy too, so they tend to not just throw things away but to instead have their old equipment decommissioned (stripped of anything proprietary, like software/firmware) and sold for cash ... yet when that happens, whomever gets charged with seeing it through often flakes out and doesn't really remove the proprietary bits. It's like companies selling off fleets of old PC inventory and the people in charge being too lazy to wipe the hard drives. Anyway, more likely than not, all the EPG Jrs that showed up were probably meant to be sold as "Ataris with some interesting useless hardware" attached -- but with the EEPROMs having supposed to have be erased. Oh well.

PS - They were only used in small SMATV MDU cable systems in the later years. In the 80s, the Jrs were used by full-size cable systems too. They were simply the less-expensive "lite" version of the EPG Sr. (supporting fewer ads, fewer capabilities, and thus having a lower monthly UV invoice). You can see one in action here: http://www.youtube.com/watch?v=3j89RFN3qps Unfortunately the video is terribly screwed up (I've seen that jumpy/jerky motion effect before: it's the result of improperly converting interlaced video to progressive), but it's still watchable.

If/when I get to the end of the atari disassembly I will try the amiga one, but I guess the coding will be an order of magnitude harder to reverse-engineer, as we think it was written in C.

Even worse: C++.

Unfortunately the megaupload link says it's unavailable ATM, I will try and download ASAP.

I've noticed this happening a lot recently. Not sure what their problem is. But often if you just hit reload ... wait 15 seconds ... then repeat, you'll soon end up being able to download the file. Anyway, only put it on Megaupload because Ari's forum restricts attachment sizes.

@Ari - Good luck with the binary packing stuff, then.

swest77 wrote:

If/when I get to the end of the atari disassembly I will try the amiga one, but I guess the coding will be an order of magnitude harder to reverse-engineer, as we think it was written in C.

Even worse: C++.

Just wondering, how do you know that?

I ask because I read that it had been written in C++ on LinkedIn, but after talking to the guy I realized that the C++ application was actually just the TV Guide On Screen channel that was acquired by Prevue Networks in the mid-1990s. Plus, that one had run on Windows (:O). So I was just wondering if you had gotten it from the same source or if you knew it was written in C++ after your consulting stint with TV Guide.

D'oh! I found that LinkedIn profile as well, and that's where I got it from. Nevermind, then!

swest77 wrote:D'oh! I found that LinkedIn profile as well, and that's where I got it from. Nevermind, then!

That's what I suspected lol
I believe it is in fact written in C.

tin wrote:Also the decomissioning report sheds some light on what I now know is called a select code (the NJJR6 bit that I so far thought was some kind of serial number)

Oh, cool! I had seen this PDF before (I think it was posted to AtariAge originally) but I hadn't seen the "select code" bit. Make sense, since the startup script of the Amiga disk I have says: "run ESQ, PGL, or LG w/ select code" (and then runs ESQ with the GAwhatever select code as an argument).

So, that number does in deed appear to be called the "select code" for sure.

BTW, that PDF's unit appears to have had two -- NNJR6 (not NJJR6, BTW), and JRMV4. (Surely just because at some point, its code was changed.) But what's interesting is that both have "JR" in them -- because the select codes I've seen in various A2000 PGs all have "A2" in them (my city's was A22050; one of the rebooting prevue guide videos on Youtube shows A22290 -- it's the video where the bottom half of the screen is solid red at the beginning of the video; and Ari's blue grid disk is GA24005).

Ari, here's something you might consider for when you write back to your UV contacts -- and for any new ones you might find. Try asking whether they might still have a copy of the printed EPG Jr./Sr./PG operator's manual sitting around that they don't want anymore. Such a thing would probably provide a wealth of information about the software's operation from a technical standpoint. I know they existed, because as a kid, I had one in my hand. (My dad worked for a phone company that was experimenting in the early-mid 1990s with fiber optic cable service -- back when all the "information superhighway" politics were in full force. Anyway, he took me on a tour of its head-end and lo and behold, there sat an Amiga 2000 in a rack with a Prevue Guide manual chained to it. (I asked if he'd let me run it through a xerox machine, but the answer was no: "the people who run that thing aren't my department, I don't wanna get in trouble," etc. D'oh! Almost had it.

Just to expand a bit more on this particular subject....

The select code on our dumped ROM is NJJR6 (btw that's where that came from) and I was initially surprised to see them so similar, hence my thought that it wasn't a unique serial number, rather a bunch of letters that indicate what the system does and does not do. perhaps some of the letters correspond to the capabilities of the box, hence Jrs and Srs are different, and, potentially cable providers. I can imagine large cable cos would want their EPGs showing the same stuff, but probably had many headends and many boxes. The title being broadcast makes sense as UV could control their boxes, they would show the correct title for the provider they leased the box to. This would require some part of the select code to be unique to the box or group of boxes though. Nonetheless, surprised to see JRMV4 which is very different!!

Note also the code in our dumped rom is actually BNJJR6($00)($00)($00)($00)($00), hence the hearts after the serial number in the atari output ($00 = a heart char in its charset). The B seems to have a special significance and is considered not part of the select code it seems, and is sometimes displayed as B:NJJR6 Perhaps Amigas were G:, just going on what we've seen on youtube ect, and what you posted. I will check the assembler code again to see if I can find where the significance of that letter comes in.

As mentioned before the addressing code allows the broadcast to contain a contain a "?" in any given char position as a wildcard for that digit, and a * if there's no need to check any further in the serial code, both of which allow updates to be broadcast to a subset of (or all) EPG boxes. Perhaps therefore the select code is structured to denote certain features enabled or not. So far the codes we've been generating (manually or with the cgi scripts) are all with the * code to address any box that's listening.