Re: [idn] Where to do form-folding?

[bill@mail.nic.nu (J. William Semich)]

An implementation of an IDN protocol is a forward-looking process - not a
presently-constrained one. The idea that there exists some constant, call
it "existing CPU resources," which should somehow constrain solutions
targeted to the (12-24 month) future is a non-starter. Given, any solution
will need to work under current CPU constraints (and other resource
constraints as well). But Moore's law is still at work. And even if you are
limiting a solution to one that reduces impact on "existing CPU resources,"
what do you take as the base? Sparc 20s? Enterprise 450s? IBM mainframes? etc.

In general, resource capabilities have expanded as the market has demanded
in order to meet the needs of implementations. 

NSI's decision to upgrade its DNS hardware instead of rewriting its
software in order to meet what was previously believed to be a memory
limitation of BIND 8.2.2 P5 is one example of resource capabilities
expanding to meet an implementation need (the rapid increase in the number
of SLDs in the .com zone).

That said, yes, it's important that the impact of an idn solution on *all*
resources be considered - not just the impact on (constantly changing) CPU
resources. 

Bill Semich

At 10:33 PM 7/24/00 +0200, Harald Tveit Alvestrand wrote:
>At 17:28 23/07/2000 -0400, RJ Atkinson wrote:
>> >General observation: If development of Internet had been constrained by
>> >existing CPU resources in the past, it would not have grown to its 
>> current level of importance throughout the world.
>>
>>Au contrair.  The development of the Internet to date HAS actually
>>been constrained by past CPU resources.  This issue is present
>>in many IETF standard protocols, from routing to applications,
>>and has regularly been considered part of the overall environment.
>
>or to put it another way: "rough consensus and running code" requires that 
>there exist CPU resources that can run the code....
>
>            Harald
>
>--
>Harald Tveit Alvestrand, alvestrand@cisco.com
>+47 41 44 29 94
>Personal email: Harald@Alvestrand.no
>
>
>