PART 1: "WHY STANDARD MODEMS MAKE LOUSY TTYs... THE 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". Why did we do this? What are the problems with using standard modems and software for deaf telephone communications? Who is NXi anyway? These issues will be discussed in the following parts: Part 1: "Why Standard Modems Make Lousy TTYs... the problems" Part 2: "Why Standard Modems Make Lousy TTYs... NXi's solutions" A third part will then be offered on the new capabilities possible in TTY communications, titled something like.. Part 3: "TTYs of the Future" PART 1....... First, some further background... in 1992 NXi received a grant from the National Institutes on Deafness and other Communications Disorders (NIDCD, an institute of NIH) to develop and improve TTY technology. NXi was founded in 1990 by three electrical engineers (one of whom is deaf) as an engineering consulting company developing advanced products for other companies. The NIDCD grant allowed us to concentrate our energies on TTY technology, and we have remained focused in this area for the last four years. Our goal has been to produce a high speed "TTY" modem suited for person-to-person "telephone" calls. But, as we looked at standard modems from this perspective, we were dismayed at their design. The deaf need a "modem telephone" which acts like a "telephone"... but high speed standard modems are very "un- telephone-like". A "TYPICAL" VOICE CALL Suppose you are a hearing person and you call someone. The person you are calling may not answer the call, but your telephone lets you immediately speak to the "wrong" person and ask to speak to the "right" person. Your call might then be placed on hold and transferred to another extension, or you might just wait for a minute of two for the right person to pick up. Once the correct person answers then communications can again occur immediately. If a line disruption or "click" occurs it might be annoying, but it won't cause your phone to hang up the call without your permission. Lastly, your telephone and the remote telephone are not going to have different configuration settings, prevent you from talking, and then abort and hang up your call! A "TYPICAL" STANDARD MODEM CALL The simple scenario above is completely foreign to modem operation. If a properly configured and "ready" computer modem does not answer your call, then your call is most likely in trouble. If a hearing person answers the modem call this is not good. First, the hearing person will hear only a "dead line" since an originating modem makes no sound. So, this person will likely just hang up. Even if this person decides it could be a modem call, his computer may not be ready and so you will hang up since you did not get a response. Perhaps you are lucky and a computer modem does answer and the connection (carrier) is established. This process takes 30-60 seconds of modem "handshaking" ("beeps & boops") while you wait. However, suppose you have reached the wrong person's computer. Your modem will not allow the call to be placed on hold and transferred to the correct person's computer! Further, any line disruption (caused by someone picking up on another extension, etc.) will likely make your modem decide to abort the call and hang up without your permission. And even if you have reached the correct modem, you may well have incompatible modem configuration settings (parity, data bits, etc.) and can't communicate anyway. If you pass all of the hurdles above your echo style may be wrong and you can't see your own typing... The obstacles above are so ridiculous they may actually seem funny (to read, but not experience), but they are among the reasons that conventional modems and communications software are the most user hostile consumer items in the world and completely unsuitable as "data telephones" for person-to-person calls. Using a computer modem for personal calls is almost impossible for ordinary mortals. This is why the deaf generally refuse to use standard modems and software as their text telephones. WHY ARE STANDARD MODEMS SO USER-HOSTILE? Modems were invented in the 1950's to allow mainframe (and later, "mini") computers to exchange computer data over phone lines. Modems reached their present form when all users were technicians wearing white coats in "glass room" computer installations. Modems have always been used for computer-to-computer calls where the answering computer was ready for the call. Think about it... this is still how modems are mostly used. People call their Internet provider (or Compuserve, AOL, etc.) and a waiting computer answers the call. Modems were never meant for person- to-person calls, or to act like a "telephone" for personal calls. THE FIRST "MODEM TELEPHONE"... THE TTY In the late 1960's three deaf engineers developed a new type of modem called a TDD or TTY. Why did they do this? What problems did they solve? First, ALL the parity, data bits, echo style, etc., etc. problems with standard modems go away with the TTY. There is only one protocol... how refreshing. But the main difference between a TTY and a standard modem is that TTYs (or people) do not have a "carrier", and standard modems do. Why did the deaf engineers do away with the "carrier"? First, what is a "carrier"? A modem carrier is a constant sound on the phone line produced by the modems on each end. A modem sends information over this carrier. The presence of the carrier is constant, and so it makes sense to say that two modems are "connected" (by the carrier). It makes sense to say the carrier will be "broken" (interrupted) if the call is put on hold. If the modem carrier is "broken", the modem has no where to go... so it hangs up (does this sound like a "telephone" ?). People do not have a "carrier" (thank goodness), so a voice call has no "carrier". There is no "connection" to "break". So, a voice call can be put on hold just fine. The deaf engineers who designed the TTY made it like a voice call and not a modem call. If a person speaks on the phone there is sound... if not, not. It's the same with a TTY. A TTY only produces sounds when you type. So, a TTY call can be put on hold and transferred to another extension and the TTY won't mind a bit. MODEM DESIGN AT NXi So, NXi looked at standard modems, voice telephones, and TTYs... what a mess! Our goal was to produce a high speed modem device compatible with both TTYs and standard modems. But, standard modems are simply not designed for person to person "telephone" calls! What is worse, standard modems are not only incompatible with TTYs, they are even generally incompatible with the 300 baud ASCII found in many TTY machines. Truly a mess. One point needs to be repeated here since there has been confusion of this... ---> standard modems will not communicate with TTYs. Further, the more we looked at the situation, the worse it appeared. Standard high speed modem connections take too long to establish (this is the modem "handshake" or "beeps and boops"). If the call is from a TTY, who wants to wait around for 30-60 seconds while the modem tries for a high speed connection before going to TTY ? And even if both sides can support high speed connections, why should one wait around for 30-60 seconds for a high speed connection to happen when TTY or 300 baud connections happen very fast and are just fine for typed conversations? ... however... a high speed connection might be needed after all if you decide to send over much information or a file... hmmmmm. Another problem... how can a high speed modem act like a telephone, never "abort" a call, and allow a call transfer since they must have a "carrier" ? And what happens if a person answers a call, but their computer is not ready? Voice callers get told to "please hold" all the time... Further, what about the communications software running on the computer? Standard communications software is designed to call "BBS's", the Internet, or other on-line services, and is simply not designed for person to person "data telephone" calls. The design of standard modems and communications software, FROM THE GROUND UP, is hostile to deaf telecommunications !!! What to do? In the Part 2 posting to deaf-l, I will discuss the solutions that the engineers at NXi have come up with to these many difficult problems. It will be interesting to see the level of interest in the deaf-l newsgroup to these issues. Tom McLaughlin Electrical Engineer NXi Communications, Inc.