Re: [apps-discuss] IETF technical plenary: the end of application protocols

[Ted Hardie <[e-mail]>]

On Mon, Mar 21, 2011 at 9:58 AM, Hannes Tschofenig

> Not really. It allows people to innovate at full speed. There is no need to go through a 5 year standardization effort for a protocol extension when you can make it available to others easily (without even letting the user to download and instal a new application).
>

So, there is an old saying: "Success has many fathers, but failure is
an orphan".  While I strongly agree that establishing where
interoperability is needed should make for an interesting plenary
discussion and is something every WG should consider, I think we have
to be pretty careful in reading any paternity tests for the pace of
innovation.

There are several mutually reinforcing network effects here.  The web
provided the opportunity for many people to become publishers as well
as consumers of network resources.  That created a pool of developers
and the opportunity for those who wanted to build tools for that pool
of developers.  That created a further explosion in interest in
creating content and new innovations in the kind of content that could
be created entirely within the web context (at one of my early
employers, we created interactive content using web forms and Sun RPC,
and the transition in technologies took a few years).

While all of this content was moving onto the web, firewall and NAT
technologies were getting widely deployed.  One of the biggest risks
was unvetted web-based content, but it could not be sequestered
without enormous outcry, so those ports were left open.  This pushed
other kinds of traffic onto those ports, even when there was no user
produced or consumable information involved.  That increased the set
of developers who understood network programming, but only from a
web-centric view point.  That, in turn, created a new body of content
and protocols built on those tools.

Network effects are very powerful, as we all know.  The availability
of one set of tools in the web environment (downloadable javascript)
has clearly had an important part in that.  But let's not confuse its
role with the overall effect.  It's particularly important that we not
ascribe to those tools the overall pace of innovation.  Developer
community size, security models, and network topology restrictions
have been as or more important.

Any of the Daves in this conversation could create an innovative
protocol to solve a problem without touching the web.  It might take a
good bit longer to deploy than one that did, but it's a set of other
constraints (treatment of downloaded standalone applications vs.
downloaded web applications, NATs, Firewalls, developer interest) that
would slow its pace.

That's important in part because new network effects are always
possible.  If we ascribe their results to one particular set of tools
we may miss the generative possibilities of other tools.

regards,

Ted Hardie