couchdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Paul Davis <>
Subject Re: Using HTTP headers
Date Thu, 19 Feb 2009 03:41:22 GMT
On Wed, Feb 18, 2009 at 9:35 PM, Noah Slater <> wrote:
> On Wed, Feb 18, 2009 at 07:49:06PM -0500, Damien Katz wrote:
>> Not that I necessarily disagree (this might be the most pragmatic way),
>> but it kind of defeats the whole point of HTTP headers. It's just taking
>> the same inputs and making them slightly harder to parse.
> What do you think the whole point of HTTP headers is?
> They are, by definition, meant to be related to the protocol.
> I don't see how full commit operations are related to the transport protocol.

My first thought was to go and look through the existing list of
headers to see if there was a non-CouchDB one that had similar
semantics. I did that for awhile and then got irritated at the level
of crap I was sifting through.

My second thought was: Are the commit semantics part of the protocol
or part of the request? I mean, are we requesting that we make a full
commit or are we making the same request but altering the protocol to
ensure a full commit?

My third thought is that we can be purists or pragmatists, but not
both. We are already not being pure in a number of respects so I find
it a bit disingenuous to claim purism on a given topic while blithely
ignoring concerns on other topics.

My fourth thought is that I have no more thoughts so I should end this email.

Paul Davis

> --
> Noah Slater,

View raw message