aBlog

From what I found at the Australian Unix Users Group (AUUG) conference a few months ago, it seems very likely that Unix will live on. Perhaps not the Unix as written by Thompson and his friends, but a version of it that is still recognisable enough to be called the original McCoy.

To be sure, Unix might have survived thus far because it was portable and popular. Back when AT&T first made it available, most Universities were freely provided with its source code. As a result, an entire generation of researchers, programmers and system administrators grew up on it and "knew no other" when they graduated. Not surprisingly, they recommended and commissioned Unix systems wherever they gained positions of power. Microsoft, incidentally seems to be copying this exact strategy to popularise what was widely referred to at AUUG as "the other operating system" -- Windows/NT, no doubt. But is this enough? Today, viability is closely connected to openness. In order to survive into the future, a system must not only be popular among a whole herd of "propeller head" graduate students who represent the future IT workforce, but it must also allow them to argue with senior management that it will inter-operate with other systems efficiently. This is openness. No longer can a business be locked into dealing with a single solutions provider and continue to be viable.

So the question is not whether Unix will survive or not, but whether Unix is open. Out of this can of worms emerges the more important question: "Which Unix?". As Jon Hall, senior manager in Digital's Unix software group pointed out correctly during his keynote, "the primary requirement for a system to be open is for its source code to be freely available." However, "anything that is free", Eric Allman of the Sendmail consortium says "is open to abuse", drawing upon the widespread abuse of the Internet and email as an example. For the Internet to survive, some restrictions must be imposed he says. He foresees the levy of a small but significant charge for each email message, thus discouraging spammers. According to Allman, even if a resource like electricity is freely available, there will always be reasons to regulate its use through charges. Can we see the same fate befalling Linux and the other free Unixes eventually? Can corporations abuse its free availability by altering the source code and then making the alterations proprietary? Hall says that this all too familiar situation with AT&T's Unix will never happen with Linux which is protected from abuse by a carefully constructed GNU Public License (GPL). It requires people who modify the system to distribute their modifications in source code too. Thus argues Hall, for a new (albeit circular) definition of Open systems -- "A system is open if and only if it can never be closed." In other words, an open system is one that is protected by the GPL, like Linux.

Is Linux then the right choice for a start-up business today? Many users have in the past shied away from this choice. Traditionally Linux has only enjoyed a great deal of popularity outside the corporate environment. But Linux advocates at AUUG argue strongly that most stated reasons against corporations going Linux stemmed from popular misconceptions, principal among them being lack of support. I was pointed to several delegates, a representative sample from a reputedly growing number, who make a living out of supporting Linux systems. Further, more and more commercial software is becoming available on the platform, what with Digital endorsing the product and all. For those undecided still, the Infoworld best product of 1996 award to Red hat Linux 4.0 should help. All this makes Linux an extremely attractive option for companies considering a move to Unix.

But what about Windows/NT? Microsoft argues that while Linux is a product of several hands around the world, NT is centrally controlled and maintained thus making it a far more reliable option with a homogenous interface on widely different architectures. However, for most people considering a move to NT, they must be able to import their existing applications and environments into it in order to successfully migrate. While almost all Unix applications can be directly compiled under Linux, Windows/NT seems to have a problem when it comes to porting programs from Unix. Tim Failes of Advanced User Systems disagrees. "NuTCRACKER from Datafocus, for example" he says "implements a Unix API under Windows/NT so Unix applications can be recompiled under NT with a minimum of code changes." But even then, it is worth remembering that NuTCRACKER and its siblings, not unlike their successors to come, are proprietary software that come with a price-tag. Admittedly, this doesn't seem to have discouraged users from going the NT way. A recent survey in August by International Data Corporation of some 200 Unix workstation users shows that 78% of the 200 respondents had purchased or were planning to purchase Windows NT workstations. And nearly 48% indicated that they had purchased Windows NT workstations instead of UNIX workstations. However, more than 32% of the NT workstations were purchased to add to existing projects, indicating a coexistence of UNIX and NT workstations in many cases (http://www.unixnt.com/9708/9708.htm).

One of the main reasons, apparently, for going NT was the apparent lack of personal productivity tools under Unix/Linux. It will be interesting to see how the trend changes as more and more such tools are commercially being made available for Linux. As the battle continues, each successive Unix conference promises to be ever more interesting to the Unix and NT enthusiasts who each have their personal agendas against or for Microsoft or the other proprietary system makers.

Dr Bob Glass, Director of the Science office at SunSoft, remarked at the UniForum NZ 97 conference that he, like most of his colleagues didn't care much whether their ideas and inventions reached the public through Microsoft or Apple, as long they did reach the public. Non-academics seldom seem to share the same indifference towards corporations. In my experience, no conference on Unix has ever been complete without at least a nominal dose of Microsoft bashing. AUUG proved to be no different. Microsoft seems to have become so formidable in their eyes that no one dared even mention the names of its products. Following faithfully in the footsteps of the first keynote speaker, Kirk McCusick, Windows and Windows/NT were continually referred to as "the other operating system". Nothing, it seemed, was more refreshing than the joy of being united in resentment against a common enemy. Were people fighting for open systems or against Microsoft? It was at times hard to tell. Its significance should be clear. The atmosphere can perhaps best be summed up in a rhetorical quote by the Linux/Open systems proponent Jon Hall from his t-shirt (http://www.readynetgo.com/shirts): In a world without fences, who needs Gates? If these people have their way, then Linux is definitely the Operating System of the future. It will be open and it will be free. Live long, learn Linux and prosper.