PART 1B: "WHY STANDARD MODEMS MAKE LOUSY TTYs... MORE PROBLEMS" BACKGROUND: NXi Communications, Inc. has now introduced the first TTY compatible high speed modem (19,200 bps data / 14,400 bps FAX), along with MS Windows software called "NexTalk". PART 1B....... Well, in Part 1 I promised to start next a "Part 2" that would discuss "solutions" to the problems mentioned in Part 1... what an optimist I am ! (You can tell this is a work in progress). This Part 1B section will mainly cover details of some of the features of TTYs and standard modems. This background is necessary for further discussions.... so here we go.... ****** It seems obvious that the 45 baud speed of TTYs (in the U.S.) is a real problem. Wouldn't it be nice if you could use your TTY to call a movie theater and get show times, or a retail store and get prices, or a government agency or business and get informative bulletins? This is not very feasible at 45 baud! For example, a six page text bulletin (25,000 characters) that takes 3 or 4 seconds at 19,200 bps (with data compression) will take an hour and 14 minutes on a TTY! (or 14 minutes at 300 baud) Yikes! Yet, all sorts of interesting possibilities could become possible in deaf telecommunications if high speed data links were made available. Some have argued that the 45 baud TTY is an anachronism and the deaf community should switch over to standard high speed modems. Unfortunately, this suggestion is misguided because standard modems are very complex to use and are simply not designed for person-to-person "data telephone" calls. This series of essays will attempt to explain why this is so. In comparing TTYs and modems, it is not so simple as "faster is better". A fast 28.8k modem is not "better" than a TTY in important respects. As was mentioned in Part 1, the "carrierless" operation of the TTY and its single "protocol" makes it very easy to use and suitable for "data telephone" and person to person calls. A TTY acts like a "telephone". The three deaf engineers who designed the first TTY knew what they were doing. At NXi we believe that a high speed modem designed for deaf telecommunications and person-to-person calls must be a combination or "synthesis" of both a TTY and a standard high speed modem. The best qualities of the TTY must be preserved, yet there is a real need for high speed connections in deaf telephone communications. So, what exactly are the problems with standard modems for deaf telecommunications? Why didn't our research team go buy the latest 28.8k modem and say, "yup, here ya go" ? Problem #1: STANDARD MODEMS ARE NOT COMPATIBLE WITH TTYS Well, right off the bat we run into this one. The deaf community uses TTYs (for good reasons), yet standard modems are not compatible with TTYs and will not "talk" to them. Periodically at NXi we get calls from people who say "I just bought a new modem, and I heard about your software that talks to TTYs, and I want to buy that software" ... (big sigh). OK, so we start explaining that our "NexTalk for Windows" software only works with our modem. "Well, why is that?" the caller demands to know.... (second sigh).... so, lets discuss this issue in some detail. STANDARD MODEMS AND TTYS It is important to realize that TTYs and standard modems have both had long, honorable, and very separate histories. TTYs were invented almost 30 years ago but trace their ancestry to the old teletype machines, and before that to telegraph machines that used a "Baudot" character encoding (you know, the things Indians cut all the time). Standard modems, on the other hand, are about 40 years old and evolved out of mainframe computer installations (Teletypes and telegraphs predate modems, so TTYs are the older tradition!) Standard modems use various character encodings, with "ASCII" being dominant in the United States. So, difference #1: TTYs use "Baudot", and standard modems use "ASCII". (If I start explaining what a "character encoding" is, I know you will snooze... e-mail me and I'll prepare a standard boring response). One other point about Baudot characters should be mentioned. Since the Baudot character code is only 5 "bits" long, there are only 32 unique Baudot characters. "Aha!" you cry, "gotcha! I can type all 26 letters of the alphabet, plus all 10 numbers, and various punctuation marks, which is more than 32!" (glad you are on your toes). Well, Baudot is very peculiar... there are two separate "tables" or "cases" of 32 characters each. If I type "I WANT 4 MUFFINS" (I must be hungry), then after the "T" the TTY has to send a "case switch" character to switch to the "numbers" (also called "figures") case table to send the "4", and then send another case switch character to switch to the "letters case" table to send the "M". TTYs do not have upper and lower case at all ("Aa" "Bb", etc.) but rather have the strange "letters" and "figures" case. The reason I mention this exceedingly boring detail is that sometimes a TTY (even ours) will get into the "wrong" table or "case" and you will see a string of numbers and punctuation marks instead of letters. Most people call this a "scramble", and any TTY software worth its salt will provide a way to fix this. So, in NexTalk if you have a TTY "scramble", you block the offending section on the screen with the mouse, choose "Unscramble TTY Text" from the menu, and presto, the numbers gobbledegook on the screen will turn into perfectly readable text (amaze your friends). I've had conversations with very smart (hearing) software programmers writing code for our hardware. I will say "so, how will you handle TTY case problems?"... "Upper or lower?" they will ask. An aside: when will mainstream communications software (Procomm, Crosstalk, etc.) ever understand enough about TTYs to support Baudot cases (letters and figures) ? I dunno... oh well, back to modem hardware issues. OK, so now that I've bored you with "Baudot" and "ASCII", watch this.... There are two other important differences between TTYs and standard modems. One, the 45 bits per second (bps) speed of a TTY matches perfectly the speed of popular teletype machines of the past, but does not match the 110, 300, 2400, 9600, 14,400, 28,800, etc. bits per second speed of standard modems. Secondly, the "modulation / demodulation" method of TTYs uses non-standard frequencies. ("Double yikes" you say, "now what is he talking about?") Well, this too is a bit nasty to explain, but a modem "MOdulates" and "DEModulates" ("MODEM", get it?) ones and zeroes to transmit and receive computer data (like typed characters), and the frequencies used by TTYs for this are not "standard". Whew... I'll try and not have too many paragraphs like that one. So, the Baudot character encoding, the 45 bps speed, and the modulation protocol of TTYs is thoroughly and completely incompatible with standard modems. By the way, to make matters even worse (they can get worse?), there is not even an official TTY "standard" to follow ! There are many, shall we say "unconventional" TTYs out there that don't really agree on how things should be done. These problems are not something the communications software running on the computer can "fix". These problems are buried in the design of modems. So, you! you people who demand we sell software to make your modem talk to TTYs... it can't be done! ****** OK, now we understand the important differences between TTYs and standard modems. Surely adding 45 bps Baudot support to a standard modem cannot be that hard, right? A TTY is a type of modem, and the TTY protocol is just one more modem "protocol", right? So, once mainstream modem manufacturers build TTY Baudot protocol support into standard modems then we're all set... Things are not that simple. Modems are suppose to figure out what the "other" (remote) modem on the other side of the phone line is capable of and connect at some mutually agreeable protocol. This is why your new 28.8k modem will (in theory) connect to old 300 or 2400 bps modems. This brings us to issues of modem detection (modem designers call this "automoding") and modem handshaking. Do you suppose standard modems support TTYs in their detection and handshaking procedures? Right. The ideal "high speed TTY modem" will be able to detect whether the remote modem (or TTY) is capable of high speed or 45 bps Baudot, and connect to anything (hey, let's add fax while we're at it). If a modem manufacturer only adds TTY Baudot support then how are you going to use the modem? If you call a standard modem and you are set to TTY your Baudot tones will "toast" this standard modem. And do you know who will be calling you IN ADVANCE so you can set your modem up to TTY or high speed? (If you do, your life is too predictable!) Simply adding TTY Baudot support to standard modems is not sufficient. TTY support must be added to the modem detection and connection procedures. Side note: these essays discuss only the 45/50 bps Baudo TTYs used in the U.S., Canada, Australia, etc. NXi's technology can be extended to support the deaf telephones used in Europe and elsewhere (EDT, DTMF devices, non- standard V.21 machines, etc.), and we will begin work on this in our upcoming V.18 efforts. Some wise fellow once said that "form follows function" (I think that guy built houses or something... ). The "function" of standard modems and communications software has always been to call a waiting and ready computer and their design or "form" reflects that. The "function" of TTYs has always been to call a person (who might or might not be waiting or ready). Engineers (that exalted species) create the "form" of a product to match its use or function, and the uses of TTYs and standard modems have always been very different. As you might suspect, this different orientation in the design of TTYs, standard modems, and communications software leads to many more "gotchas" and problems to be discussed before the "combination" of a TTY and a high speed modem can be envisioned. ***** These essays seem to be growing larger than I planned. The next posting will be on (truly gruesome) details of modem detection and handshaking that are barriers to high speed modems being used in deaf telephony and person to person "data telephone" calls.