Peter Flynn writes in response to Gregory Murphy's quote of Jim Clark's
comment on "open standards":
> If this commitment [..., Netscape is commited to the idea of "open
standards" > ...] is true, it marks a major departure from past Netscape >
In the software industry, open standards means nothing more than
non-proprietary. This has nothing to do with what seems to me to be
central in all this discussion -- who creates standards? In the software
industry, software companies create preemptive standards, customers
create de-facto standards, and standards organizations create public
Preemptive standards are usually, but not always proprietary.
Microsoft's Licensing (LAPI) standards will not solve the issues brought
to us by superdistribution, because the LAPI standard was only put
together to handle the SBA's problems with software piracy. This will
leave us with a crippled standard when superdistribution is finally
feasible. These standard's are usually created in the absence of a
market. The economics for these products are extremely risky. Create it
and they might buy. Most (over 95%) software companies fail, because
they believe the market exits when it doesn't.
De-facto standards usually involve the selection of a previously
preemptive standard. The economics here focuses on satisfying demand.
The company created it and the customers are buying it. Of course, the
competing standards are heaped on the trash bin.
De-facto standards usually follow the economics of increasing returns.
In the Netscape vs Microsoft battle, Netscape will probably lose,
because Microsoft can afford to tie content to its browser. But, then
the battle over open standards will have been lost, because the
technology that forces sites to create Microsoft Explorer content is
proprietary. If public standards were followed, I wouldn't need another
browser, but hey I can get twenty browsers for free.
Public standards, like SGML and CALs, are usually created well in
advance of any products. Public standards do become software standards,
because the standard creates a market for the product. The existence of
a preexisting market helps software companies, because all they have to
do is comply with the standard. These companies do not have to spend
lots of money creating a market. The economics involved is pretty
straight forward. Create it and they will buy it.
The TEI standard will create a market. Before TEI, few people actually
grasp what this kind of standard would mean. Now TEI is a known
commodity within the TEI community. People outside TEI still don't get
SGML let alone TEI. And, as long as TEI users insist on compliance with
the TEI standard, software companies will create the software without
trying to go the preemptive route. They can make enough money just
following the path created by the TEI standard. But, the TEI community
must not compromise. If a product deviates from the standard, don't buy
it. The first software company that succeeds in getting market
acceptance for a product that puts out non-standard TEI will kill TEI.
> In Mr Clark's speech at CERN earlier this year, when made it clear that he
> considered standards were something to be formed by companies to give
> them a competitive advantage, not something to be discussed publicly,
> let alone decided publicly.
He is absolutely right as far as software company created standards.
> [And quote of Jim Clark again]
> "A short while later I asked Mr. Clark if, given his impression of the
> intractability of the IETF and his commitment to open standards...
Commitment to an open standard does not imply a commitment to a public
standard. Open means that the standard is published and can be used by
outside interests usually without license. Public standards are
inherently open. CORBA is open. OLE is proprietary. Although, you could
probably learn enough about OLE from open sources, it is only supported
on MS Windows. Open standards are usually thought of as platform
independent, or not limited to the wintel platform.
It should be noted that commentator's usually ascribe the replacement of
proprietary IBM systems by the PC to the PCs openness. The IBM PS1 line
was closed. The amazing thing is how closed the PC has become.
Openness will win in the long run -- ASCII, the ultimate open system is
still with us -- but, how long is the long run? And, how inefficient will
it be until then?
David Locke [log in to unmask]
Engineering - Drilling Information Products
200 Gillingham Ln. Voice: 713 275 4722
Sugar Land, TX 77487 Fax: 713 275 8098