BABBA issue # 8 - October 1993, pages 12-30
Do you need to use your computer for more than one thing at once? I do, and apparently there are millions more like me, as demonstrated by the sales of DesqView, OS/2, Windows, etc...
What do most graphical multitasking Operating Systems have in common? High system requirements and slow performance. OS/2 (v2.1) for example needs 8 MB of memory and 35 MB of disk space just to run a few small or medium size applications. For major applications to run well, you need even more memory and disk space.
The AGD (Association of GEOS Ensemble 2.0 Developers) is currently developing commercial and shareware software for Ensemble 2.0, some of which will be available as shareware on America Online. (Keyword: GEOWORKS, search file areas for "AGD")
A growing amount of Ensemble 2.0 documents, drivers and utilities are arriving on local BBSs. The CyberPort BBS (408) 227-1635 (BABBA ZONE 3) has lots of Ensemble 2.0 files. Some examples: GeoFaxes (from the fax on demand information system provided by GeoWorks), new printer drivers (to improve printing speed), clip-art, icons, and much more.
A great feature of Ensemble is that most of the applications are scalable. For example, if I set GeoWrite to user level one, I have what amounts to little more than a glorified text editor. If I set it to user level four, I have a program that is adequate for desktop publishing. All the Ensemble major applications and most of the included programs and utilities are scalable. Using the Complete Communicator fax board, Ensemble 2.0 supports faxing. You install the fax driver in a fashion similar to that of a printer or plotter. You can then send a fax as easily as you would print.
Ensemble 2.0 cannot run a DOS application in a window. It does not support OLE (Object Linking and Embedding). It will not work with DesqView. It cannot task switch DOS applications without a task switcher such as DR-DOS v6 TaskMax. Ensemble 2.0 does, however, support TaskMax. DOS applications run in Ensemble while under TaskMax. The DOS applications are run in their own separate memory area, the same way programs started from TaskMax do. The DOS applications are automatically listed in Ensemble 2.0's Express Menu. This allows you to start a DOS application without interrupting your work.
Ensemble 2.0 still does not have 256-color drivers. They were under development, but were dropped. They had too many last minute problems for this current release (I was a beta tester). Ensemble is also lacking sound support. It currently only supports the PC Speaker and it doesn't take full advantage of what the PC speaker can do.
GeoWorks Ensemble 2.0 offers a complete productivity environment without requiring a powerful and expensive computer. I hope my review helps you determine whether GeoWorks Ensemble 2.0 is for you.
The BBSopoly game is designed for the entire family. It is available for multi-node BBSs that use Galacticomm's The Major BBS software. Game play is very similar to the conventional Monopoly board game. To begin, you must have at least one other BBS caller online to play with. After you have an opponent, type START to begin a game. Type ROLL, and the game will roll two, six sided dice.
You are then asked if you wish to purchase the property you have landed on. You have two choices, you may either buy the property for the price it is worth, or you can type N for NO, I do not wish to buy the property. You will then have to bid with the other players, to see who will own the property. Once all the properties are owned, the game starts to get interesting!
When you run out of money to buy a property, or you owe someone money, you may MORTGAGE a property and get so many dollars. All the math in the game is done for you. When you build your wealth back up to par, you may UNmortgage your properties and people will start paying money when they land on the spaces you own.
The only difficult part of the game is when you trade property cards with other players. First you determine which card(s) you wish to trade. Then you send a message to your opponent, that you wish to trade X property for Y property, and you will throw in Z amount of cash. Then your opponent must decide whether or not they want to risk the chance of giving you a monopoly, and losing the game.
Like the regular monopoly game, you can buy add-one for your land. In the original game you could buy houses and motels. In BBSopoly for The Major BBS, you buy 2400 and 9600 baud phone lines!
If your (out of warranty) external modem stops working, or becomes "confused", consider checking the capacitors in the power supply section. This applies to external modems - because internal modems get their power from the computer.
Modem manufacturers use the smallest and cheapest power supply filter capacitors they consider adequate for the job. In this article we use Boca brand external modems as an example, because they use particularly small capacitors. Boca modems are fine, but if your modem fails, this article identifies a likely cause. This advice can be applied to many modem brands.
Boca uses ultra-small "car stereo" electrolytic 1,000 uf (1,000 microfarad), 16-volt capacitors that are just too small physically for the stressful job they have to do. The extra small capacitors are used because of the circuit board layout. These tiny capacitors are not designed for the high ripple current present in the modem's half-wave power supply circuit. They quickly "dry up" or leak. Although a higher voltage rating and capacitance value couldn't hurt, the same values in a normal sized capacitor case would probably be adequate.
Of the six capacitors used in the power supply, the two that go bad are situated with enough extra space to fit "real sized" capacitors in. If you know your electronics and know how to solder, you can save by fixing your modem yourself. If you don't know what a capacitor is, or how to work on circuit boards, do not attempt the repair discussed here. Let the modem manufacturer, your favorite technician, or your favorite computer service shop do the repair.
The old capacitors need to be carefully unsoldered, making sure the circuit board holes are free of solder. The new capacitors are then soldered in, observing their polarity and keeping in mind the case will have to be put back on. Always try to get the manufacturer to repair it first. Replacing these capacitors will not always fix the modem. Think of this proposed fix as an option to just throwing a failed modem away.
We all know that modems are getting faster and cheaper. Rumor has it that they are getting better too. Article after article in various magazines give glowing reviews to one or another of the latest crop of modems. You rarely read anything bad about any of them, which means they are all wonderful, right?
Reality, however, paints a different picture. While modems are gaining features, functions, and speed, they are also getting more complex. At the same time, modems are becoming commodities in the computer world, with rapidly falling prices. New modems are entering the market almost every day, and new versions of firmware for existing modems are being released even faster.
Unfortunately, one area of the design cycle that is apparently being neglected is design verification and testing. The result of this neglect is that among the current crop of V.32bis modems, I have been unable to find one that did not have problems. In most cases, I consider the problems I have found to be fatal. With my limited resources and the huge number of modems available, I have not been able to test every modem there is, although I have been testing modems for 7 years.
If a modem cannot connect on ideal lines to its own kind, why bother with further testing of performance under non-ideal conditions? While the repeated connections are being made, count the number of times a command was given to a modem where the modem ignored it or executed incorrectly. Having a modem ignore a command, for example in a setup string, can really mess up the control over the modem.
The next test is to pass some data through the modem and see if it gets through correctly. Since the connections are noise free, there should be no data errors in even the non-error-corrected modes. Any errors seen are pretty much generated within the modems. You should, by definition, never see data errors in error-corrected modes under any condition.
Since you are passing data through the modems, you might as well pass it at the highest interface speed the modem supports. The modem should be able to use hardware flow control to regulate the flow of data without losing anything. Since all the popular connections are defined as being full-duplex emulations (meaning data can flow in both directions at the same time) perhaps you should test that too.
For a complete test, the data passed through the modems should test the error checking and data compression features of the modem. You should use both pseudo-random data and compressible data with repeating sequences. Since you are passing this data as fast as the modem can transmit it, you might as well time the transfer to see how fast the modem really goes.
Of course, before you trust any of the information you collect like this, qualify your test equipment by performing these tests over a null-modem cable to verify that the test system itself can keep up with the data flow. It's embarrassing to have an alleged bug be attributed to your test bed rather than the modem you were testing.
This is the type of testing I have been performing on modems for over 7 years. The results are disheartening. I believe that if you choose to ignore a flaw, judging it to be unimportant, it will sooner or later sneak around and bite you from behind. Most bugs manifest themselves in multiple ways.
For instance, you may decide to ignore data errors seen in the V.32bis non-error correcting mode. After all, that is not your intended use for the modem, and the error correcting modes don't generate errors. But that bug will still live in the connection under the error checking protocol, causing re-transmissions of packets and re-negotiations of the modems, resulting in a lower throughput than you should get. This is one reason why some modems are actually faster with a 9600 baud connection than they are with a 14400 baud connection.
I also firmly believe that there is no such thing as a modem that I cannot expose flaws in. This is all right as long as the flaws are minor. An example would be an occasional failed modem connection in one of the lesser-used emulations. Unfortunately, most of the flaws I have seen have been real killers. I recently had a chance to test two popular modems that you would think, from the logo, would be first-class performers. I shrink from identifying the manufacturers here because they have not been given a chance to respond to my test results.
The dialing modem hung up the call and could be reset, but the answering modem stayed off-hook. Nothing I could do to the serial port of the hung modem would get a response. I tried different interface speeds, different connection types, and changing which modem dialed and which answered. The result was always the same. The answering modem locked up, and I was never able to complete any full-duplex transfer in an error-corrected mode. Would you trust that modem for remote unattended operation? I wouldn't.
I have seen modems that generated data errors in the error-correcting modes. I have seen modems that got confused in the data compressing modes during full-duplex transfers and dropped carrier. Or they got stuck with all data stopped and sat there, connected, unable to transmit a single byte for days! (I left them testing on a Friday night. I came back on Monday morning finding them still connected.)
When preliminary testing shows so many errors, there is no reason to test the various impairments a modem should be able to handle. Or test the ability of a modem to detect a ring through the range of ring voltages and frequencies. I have even seen modems fail to connect to anything. These problems were not in prototypes but in off-the-shelf products.
2) Time:
Modems are very complex these days
and testing all the features of a modem takes
a long time. Doing a complete test and
verification (starting with the above list)
requires many months of testing, delaying
the release of the modem to production.
Most companies are unable to suffer
this delay. The world market is very competitive.
A company must race to get a new
product to market while there is still a
profit opportunity.
3) Money:
It costs money to hire good, full time modem
testers, but that is the only way to do the job.
Companies also need expensive equipment to
work with, and floor space for a lab. Some of
the testers must also be expensive programmers
to generate and maintain automated test
programs like the one I use. All this costs
money and brings no tangible benefits
that can be entered on a spread sheet.
Customer dissatisfaction impacts
profitability next quarter, while company
management is primarily
concerned with the current quarter.
4) Understanding:
The people performing the testing, their supervisors,
and the engineers that repair the bugs must
understand what to test, why it must be tested,
and believe in quality first.
Due to time-to-market constraints, quality assurance
is often seen as a necessary evil at best. Often,
reports of serious bugs are met by the powers-that-be
with denial, (it's not the modem, you are doing something
wrong), excuses, (we don't have time for this, the
problem is too hard to fix), or down-play, (that's not
important, the customer will never see that).
I know cases where engineers have downplayed the importance of a particular bug, while at the same time, someone else in the same company was trying to actually use the modem and was tearing their hair out over that very same bug.
1 ) The manufacturer assumes the role of an unjustly injured party. "Obviously the guy doing the test didn't understand the modem. He had the wrong setup and did the wrong thing." And sometimes the manufacturer is right.
2) The manufacturer pulls advertising from the magazine that carried the article, or refuses to pay for the full-page ad that ran opposite the bad review.
3) The magazine feels the loss of revenue and warns the testers to only write good reviews.
The end result is that you usually never read real bad stuff about a modem. If the modem almost never connects, the tester writes about how pretty the LEDs are. And you, dear reader, wind up buying nonfunctional "Super BarfData II Plus" modems with pretty LEDs.
V.32 and V.22bis are modulation techniques, which control the transfer of data between two modems. V.32 defines the modulation technique at 9600 bps while V.22bis defines 2400 bps modulation. Modulation/demodulation (modem) is simply the process of changing digital bits of information in your computer to analog tones for use on phone lines and vice versa.
V.42bis and MNP5 are rival hardware based protocols that handle error correction and data compression. Error checking is performed and data is re-transmitted if necessary. The compression algorithm boosts transmission speed by assigning symbols to common words in the file and transferring the symbol instead of the actual word.
We are often asked "Can you recommend a commercial service or BBS that offers Internet?" No regular BBS can match a commercial system for its advanced Internet capabilities.
Perhaps the most desirable feature of the Internet
is the universal mailbox. Your Internet address
is unique and works throughout the world.
Most people just need a universal mailbox or access to a
few Internet newsgroup conferences. For those people,
a BBS works just fine. Many BBSs offer Internet
mailbox support.
This month we identify some local Waffle
BBSs that offer Internet mailboxes for free,
or for a very modest charge:
Dark Side of the Moon,
Mookie's Place,
Luna City BBS,
Micro-Medic,
Shakala BBS,
Spectrox Systems,
Telesoft,
The Canned Ham,
The Enclave BBS,
The Land of Garg,
TheRat
Waffle BBS software isn't the easiest to figure out for the novice, but neither are commercial UNIX-based Internet providers. If you use their systems, please support the Sysops.
One of the first communication programs for OS/2 was HyperAccess/5 produced by the Hilgraeve company. The product has been continually improved, and recently Hilgraeve has entered the Windows market. Ease of installation and use are the first two things you notice. I have been reviewing several Windows communications packages over the last year. With the exception of ProComm Plus for Windows, all seem geared toward the corporate environment. HyperAccess is for everyone. This is a program for the rest of us.
It has a fully customizable phone-book that can be viewed as a detailed, statistical, or in an icon-based format. You can sort the phone-book 5 ways, and store and display them in groups. If you don't like the phone number in column 2, just drag column 2 over column 4 and they will switch. You can re-size buttons, add buttons, etc. You can select a different icon for each entry from the dozens supplied, or use your own creations.
To make a call, you can either double- click on an icon, or click on several, and have multiple calls placed. There are options on how you wish the calls placed. You can even set up and get callbacks from remote host computers automatically.
HyperAccess supports hundreds of modems. The generic or 'user-defined' modem initialization string features make it easy to set up an unlisted modem. It supports all the popular protocols and terminal emulations. It includes ready-to-use scripts to connect to CIS, MCI, and even some popular BBS programs. HyperAccess also can learn your logon sequence and use it the next time you place a call.
The Status Bar is located on the bottom of the screen and shows you the important information you want to see. Using the Launch Pad you can create a message on or off-line, and then send the completed message. If you have several modems, HyperAccess will allow you to connect to all of them at the same time. Each session has its own window, settings, etc.
Receiving and sending files was never easier. You drag-and-drop selected files for transfer. You can even save them for a future upload. The receive window shows you more information than the average transfer window. You see the number of files you are receiving, the total size and time, as well as how each file is progressing.
Another HyperAccess feature is the built-in virus filter. It scans for hundreds of virus types while you are downloading files. This happens without any loss of throughput. The same applies when using it in the background. I was pleased with the smoothness and quickness of the program no matter what else I was doing.
HyperAccess's Host mode answers calls, transmits files, allows users to leave you messages, and uses password protection to allow limited or unlimited access to your computer files. It will even allow you to dial out while the host mode is waiting for a call. You do not have to terminate the host mode. If you have several lines or modems, HyperAccess's Host will answer as many as Windows will support.
The only annoyance I found was in the way the program indicated file transfer completions. The only sound emitted is a relatively quiet beep that is easy to miss. Of the many Windows communication programs I have reviewed, HyperAccess is one of the best. HyperAccess is an excellent and well-done package. The wait was worth it. Hilgraeve offers a 60-day satisfaction guarantee return policy.
Page 17 had ads for Western Hemispherics Technology and
Shareware Express.
Page 29 was a full-page ad for PARS International Computer (www.pars.com)
Page 30 (back cover) was a full-page ad for
TeleText Communications.
End of Issue 8. Go back, or to
Issue 9, or to
Mark's home page.