Anecdotes from The Village

In writing about Global Village Communication over the last few of these posts I was reminded of a few stories that didn’t quite fit into the other narratives, so I figured I would collect a few of them here before I move on to the next stop in my telephony career.

  • The Two Logos

A lot of thought went into the original Global Village name and logo.  Rick, our key founder, talked about spending a couple of months on just that, having made the the mistake of taking name and logo too lightly with his previous company.  He had delivered a FORTRAN compiler with a company he had named Palo Alto Shipping.  That is another name that failed to make the transition to Google search.

However, when the VCs brought in a new executive team to bring the company public, they didn’t like the old logo and had it redone with more simple shapes and brighter colors.

Old logo vs. New logo

Those are from my old business cards, which I still had in a drawer.  They also ditched the sans-serif font and changed the orientation of the business cards… the originals were vertical… and generally made everything “more normal.”  Those of us who had been around for a while referred to the new logo as the “Kid Pix logo” as it looked like something made with that youth oriented graphics tool.  It was also a jab technical shallowness of our new corporate leadership, who were very much about the trappings of success and not much about the actual mechanics of it.

  • The Domain Name

In 1993 Global Village first got itself connected up to the internet.  The initial connection for the growing company was a 56K leased line, which served all of us.  The thing was, the internet service provider/registrar we went with could only support domain names up to 12 letters in length and, if you count on your fingers and the toes of one foot you will see that Global Village has 13 letters in it.

Ah, but the company name was Global Village Communication!  Why not just go with “gvc.com” for a domain?

The problem was that, at the time, there was a Taiwanese company operating under the name GVC that also made modems, and they had grabbed the domain.

A GVC Super Modem!

That company was also always confused with us because many people would refer to us by our initials and… well, there was that other company.  I don’t know what happened to them, but you can find their old modems on eBay the way you can find Global Village’s.

Anyway, the decision was made, our domain was to be “globalvillag.com” and it remained that way for about two years.

A clip from one of my 1994 business cards

This, of course, was both annoying and mildly embarrassing.  At some point in 1995 we ended up changing our internet connection and, along the way, were able to obtain the domain “globalvillage.com” as part of that.

The press release that went out with this had the headline “Global Village Buys a Vowel.”

  • Always Test or Something

Rockwell, the manufacturer who made the chipset we used in the PowerPort Bronze, updated that chipset and told all of its customers that after a certain batch we would be getting this revised chipset.  It was pin compatible and had no changed to its functionality.  They likely had just fixed a couple of firmware bugs or cost reduced the design somehow.  This happens all the time, so it was no big deal.

So when QA heard about this and asked for a unit with the updated chipset to smoke test before we started shipping them, then then VP of engineering rather sarcastically told us there was no need, it was all pin compatible and please stay in our lane.

One thing that did change was the checksum that the chipset returned when queried, and that was kind of important.

Our software used the checksum to determine which modem was installed in the PowerBook.  This was also our low effort copy protection scheme.  Since the main value of our product was viewed as the software… we were often asked to license our fax software as it was extremely fast and easy to used compared to the competition… this copy protection was viewed as somewhat important, largely because there were only two major modem chipset manufacturers in the market, Rockwell and AT&T, and we used both, so our software absolutely would work on all of our competitors hardware.

So when the new chipset hit our production line and started appearing in the field, tech support started getting a lot of calls about our software not recognizing their brand new PowerPort Bronze modem in their new PowerBook laptop.  Tech support went to QA, which didn’t have the new modem so couldn’t figure out the problem.  QA went to the software dev team, which also didn’t have a new modem and gave QA and support the usual “works for me” response.

Eventually one the hardware team passed by and heard us, asked a couple of questions, then brought up the new chipset thing and produced one of the new build modems.  Once installed the problem was verified and obvious; the software didn’t recognize the checksum and so rejected the modem.

A version of the software that worked with the new checksum was rolled up in a few minutes, but how to get it out to people who needed it.  This event predated the world wide web.  The preferred method of software distribution at this point was floppy disk.

So floppies were made and mailed to customers.  Very angry customers… who all seemed to be in the 212 area code… got a disk sent overnight via FedEx.  Local customers were told they could drop by the office for a floppy.  That hit the immediate need, but there were now hundreds of PowerPort Bronze boxes in the sales channel with the wrong software.  So there was a scramble to find and get an updated floppy to Apple authorized retailers, which is where most of the installs were happening.

It was one of those things where for months and months afterwards support would get a call because somebody bought one of those modems and the software wouldn’t work.  Tech support knew the issue at least, and there was a process, but it added overhead to every PowerPort Bronze sale.

Afterwards, every time hardware said something didn’t need to be tested… and they were pretty shy about saying that for a long time… they were reminded of the H mod PowerPort Bronze modems and all the trouble that caused.

I have used this as one of my tales for new people in QA and development.  If you think you need to check something, insist on being allowed to do so.

  • The Devil is in The Details

The flip side of the above story.  One of the… shall we say, less technically adept individuals in tech support… came up to QA with a problem.  A customer had called up complaining that they couldn’t fax from QuarkXPress (QX).

They came up to QA to ask us to figure it out and, fortunately for me, they went to Terry, two cubes over, so it wasn’t directly my problem, though we all worked together to figure this out.

The first question was whether one could reasonably expect to fax from QX.  It was a PostScript focused design application and faxing was strictly a QuickDraw experience… the difference between printing to a PostScript enabled laser printer and printing to an old dot matrix printer.  The former prints and scales based on the formulae of your output while the latter just prints dot for dot whatever is on the page or screen.

That had go back to the customer and they responded that QX did in fact have a QuickDraw print option that one could print things out for samples or prototypes.

That confirmed, next up was getting our hands on a copy of QX in order to try it ourselves.  This was difficult because the founders of Quark, Inc were extremely paranoid about copy protection and people stealing their very expensive software.  Quark was the sort of place where the floppies were removed from everybody’s Mac to ensure nobody was stealing software.

After quite a bit of work a demo copy was obtained from Quark with a limited duration license that could only be installed on one machine.  Terry setup it up, loaded up a sample document, and it faxed out just fine.

We reported that back to support, the person there having been asking us for updates in a very testy fashion for weeks, and they went back to the customer.  The customer sent their document, which was tried, and it faxed just fine.  So we went through the routine of making sure all the versions lined up and we were using the same modem and the same MacOS patches and whatever.  It still worked.

Terry went back to support and asked them to go through the steps, but the support person was adamant, the customer was just faxing and it was throwing an error.  That was all they kept saying through the whole process.

So we got on a call directly with the customer and had them walk us through the problem.  They opened their document, held down the control key to fax rather than print, selected fax, entered a page range… and there we said, “What?  Page range?”  Nobody had mentioned that to us.

We asked if they had mentioned that to the support person.  They said they had every time.  We went to the support person and asked if selecting a page range was something they had heard previously.  They said something along the lines of, “Oh yeah, they select a page range every time.  But that wasn’t important.  The failure to fax was important.”

Except, of course, the page range selection was the critical bit of information we had not been given, literally for months at this point.

QX, being a high end, sophisticated design tool, did not use the standard MacOS print dialog.  It had its own special print dialog with many more options and, among other things, it remapped many of the field to meet the needs of the product.  The page range entered into that dialog was being delivered to us with a lot of extra data that we did not expect, so we sent an error back to the app which just reported it as a printer error.  It was “PC Load Letter” in its own special way.

So we came back and said, “Don’t enter a page range and it will work fine” and the customer said, “That works, but why didn’t you tell me this five months ago?” and it was very hard not to just throw the tech support person under the bus.

A couple of years later we ran into a similar issue when Word 6.0 came out and they too remapped the default print dialog to allow people to print “sections” of a document, a new concept they invented, which of course broke us as well.  Page ranges still work though, and support wasn’t happy when we told them people couldn’t fax with the new feature, but at least it didn’t take us five months to figure it out because we had learned not to trust support to provide relevant details.

  • Ouagadougou

I had the word “Ouagadougou” written on the whiteboard in my cube at one point.  It is the name of the capital of Burkina Faso.  I had it written there mostly as a goof, an interesting name with a lot of vowels.

Our new CEO, Neil, walked by my cube one day… a rare moment to see him over in engineering… and he saw the word and said, “Ouagadougou is the capital of Upper Volta!” with an obvious amount of pride.  I, being young and ready to initiate career limiting moves on a whim, immediately corrected him and said it was the capital of Burkina Faso.

He persisted, stating firmly that it was the capital of Upper Volta because he had a big world map on his wall at home so he knew.  Once again, apparently keen to lecture somebody who could fire me, said that Upper Volta had changed its name to Burkina Faso nearly a decade back and that he should see to getting an updated map.

He was not having it.  The map said Upper Volta and it was a National Geographic map or some such, so it wasn’t likely to be incorrect.

Finally having sensed I might have ventured into perilous waters, I suggested we look this up later as we both had things to do.  This was surprisingly mature of me in hindsight, because I am annoyingly persistent when absolutely sure I am correct on a point of fact.  If I am not hedging with something like “I think…” I am going to stand my ground.

We went our separate ways and, a couple of weeks later, when I bumped into him in the lobby coming and going, he stopped me to say that his map had a publish date of 1980 and allowed that perhaps the country might have changes its name since.  I hadn’t forgotten about the whole thing, but I figured he would, so it was a bit of a surprise.  It was probably the closest I ever saw a CEO come to admitting being wrong about something.

  • The Case of Soda

Our QA group was pretty determined to get a list of changes from the software dev team every time they gave us a new release.  The devs, on the other hand, had to be nagged a bit to put together release notes.  It was the early 90s and every coder was a cowboy in those days and their code submission entries carried barely comprehensible comments, nothing you could automate a release document out of.

One of the developers was annoyed by having to put together these notes with each release because he didn’t think we read them.  So he put at the bottom of one he wrote that he would buy a case of soda for the first person to read down that far and come and talk to him.  My cube neighbor Geoff beat me to it.  The devs kept writing the release notes with builds.

  • The Color Blind Cable Guy

With the first Performa 2400bps modems there was a request by Apple to use a particular vendor to assemble them.  For tax reasons, there was some benefit if we used an assembly facility based on a native American reservation down in New Mexico rather that our usual local assembly.  We worked it out and they setup and assembled all of the early 2400bps modems.  They did a good job, were on time and within budget.

But there was a quality control issue.  An unlikely number of modems were not functional when the customers setup their system.  They called Apple, who replace the modem, but Apple saw this was happening a lot and came to use to find out what was going on.  Our hardware guys, led by our hardware design lead, who was of partial native blood himself, went down to inspect what was going.  That meant flying into Albuquerque, driving for four hours or so to get to the reservation, then spending a day going down the line to check how things were being done.  Jim, the hardware lead, said he got on well with the team there saying that they were of the same temperament as he and they too wanted to get to the source of the issue.

They finally got to the last stage of the process where they were putting the modems in their plastic cases and soldering the serial/power cable to the circuit board when Jim saw the person doing the soldering attaching the wires wrong.  The wires were, by power cord convention of the time, black, red, and green.  It turned out that the guy doing the soldering was red/green colorblind and had just been guessing.  There was some embarrassment at this revelation, but they swapped him to another part of the line and had somebody who could distinguish the cable colors take over.  The quality problems disappeared going forward.

  • Supra Invaders

Finally, a reminder of how we got under the skin of some of our competitors.  When your competitors start trying to cast you as the villain, even jokingly, you know you are doing well.

At a trade show in 1995 a couple of our reps were walking by the Supra booth.  Supra Corporation was one of our key competitors in the Mac desktop modem market, and in their both on their demo machines they had a little Space Invaders knock off came called Supra Invaders.

It featured a Supra modem at the bottom of the screen shooting Global Village logos that flew around the top of the screen dropping bombs, all over a background featuring a wrecked Teleport Platinum modem.  I still have a copy of it on my old PowerBook 190cs.  Unfortunately, I cannot get that unit hooked up to the internet to move screen shots. It has a floppy drive, but only a MacOS formatted one, and none of my other machines has had a floppy drive for more than a decade.  So here is a horrible phone picture of it.

Supra Invaders

Here is the “about” info for the app.

About Supra Invaders

Of course, our reps got a copy from them and it was all over our office within a week of them getting back.  And I guess it still survives today, even if it is effectively locked on my old laptop.

Supra itself ended up being bought by Diamond Multimedia that went on to make early MP3 players.  I had a Diamond Rio 500 way way back in the day before the iPod came to Windows.

I think that about wraps up Global Village.  There were other dumb things that happened, such as my boss sending me to a network training class focused on IBM’s token ring architecture, which was on its way out and ended up being just so much trivia in the back of my brain.  And there is one more post I am saving for later, for a 30th anniversary of an event.  But the next stop will be the big island, away from modems and into other aspects of telephony.

The tales so far:

6 thoughts on “Anecdotes from The Village

  1. Archey

    The hardware change and QuarkXpress stories really hit home with me. I have so many stories like that from when I was in support and did light QA for one job.

    I suppose this was early on and there wasn’t the institutional knowledge we have nowadays, but there are so many aspects that are so common they have become cliches. “Works on my machine!” “That issue can’t be related.” “We don’t need to test that, it didn’t change.”

    Like

    Reply
  2. Wilhelm Arcturus Post author

    @Archey – I have a lot of sympathy for QA teams. Back during that PowerPort Bronze issue there were literally just 3 hardware configurations and three possible modems SKUs. We introduced a fourth and fell over. And, in hindsight, that was such a small texting matrix compared to what I would end up facing in my career, even acknowledging that the 3×3 grid of configurations is deceptive.

    That grid doesn’t even get into the number of possible software faxing situations that existed. We couldn’t test all of that. And then there are the other possible fax machines we had to connect to. There was a line of Ricoh fax machines that were wildly out of spec so we couldn’t connect to them because we took the Bellcore fax spec as gospel. We had to go buy one of the failing units and then alter our code so that we would connect to machines that couldn’t get simple timing right.

    And, like I said, even with all of that, the testing matrix wasn’t so bad compare to things I would experience later, and where I would work for VPs who would lecture us that we couldn’t take the time to “black out all the squares” on the possible test grid but who would also yell at us if we didn’t test the right squares… “right” being defined as the squares where there were problems we couldn’t possibly know about unless we tested.

    Then there is the fact that almost every critical bug I have ever found in my life… leaving out those that should have been caught (but were not) by the most rudimentary unit test… were because I inevitably start coloring outside of the lines and immediately find that the world beyond the happy path is a jungle.

    Like

    Reply
  3. Tashtari

    I’m enjoying reading your anecdotes about Global Village, sounds like wild times.

    I wonder, do you remember any of the technical details of the TelePort ADB modems? As the only devices I’m aware of that use ADB category 0x5 (low-speed serial devices), I find them kind of fascinating, and I’m trying to pick apart the details of their protocol. They seem to be an 8051 of some kind hooked up to a Rockwell modem chip, so you’d think the microcontroller would be mainly concerned with schlepping data around, but there are so many mysterious bits in the various ADB registers and status messages, added to the fact that there’s a whopping 8K SRAM on board, much bigger than you’d think would be needed for just circular queues, that I have to wonder what complexity lies beneath the surface…

    Like

    Reply
  4. Wilhelm Arcturus Post author

    @Tashtari – The only thing I remember these days… I never actually owned one of the Teleport ADB modems so do not have one stashed in a drawer to pull apart… is that while there was theoretically a lot of bandwidth to go around on ADB, the polling rate of the chip on the Mac motherboard put an effective limit of under 10Kbps on the data rate.

    That was no big deal for bi-directional 2400bps modem traffic, but in some testing I did at one point in the lab at GV I found that while it the Teleport ADB supported the v.29 standard, most of the time it had to negotiate down from 9,600bps to 7,200bps because ADB was throttling throughput.

    Like

    Reply
  5. Anonymous

    New comment field on mobile? We’ll see if this works. Doesn’t even have a field for me to enter a name!

    I’m too young to have any memory of dealing with modems, but I just wanted to say that I’ve enjoyed this series, Wilhelm. It’s always fun to see how the more things change, the more they stay the same. Especially as I’m getting into making a career change into Tech (just starting my first job as a data analyst).

    Seraphgrim

    Like

    Reply

Voice your opinion... but be nice about it...