Media Types 'text/csv' and 'text/comma-separated-values'
duerst at w3.org
Fri Feb 4 02:56:49 CET 2005
As far as I know, there is no convention that would force either
text/comma-separated-values or text/cvs. My suggestion is to
pick one, and have implementations converge on it. I'd personally
prefer text/cvs, just because it's shorter.
At 08:08 05/02/04, Yakov Shafranovich wrote:
>The reason why I put in an additional 'text/comma-separated-values' type
>is because there is an existing registration for
>"text/tabbed-separated-values" and I thought that the CSV one should
>correspond to the existing one. However, its too long for practical use
>which is why I included "text/csv" as well. In practice I have seen both
>In truth, I don't know enough about the historical conventions for MIME
>types to judge whether an existing registration for
>"tabbed-separated-values" should mandate "comma-separated-values". I
>assumed that there is a convention, but if in fact I am wrong, I will be
>happy to adjust the draft to target one type.
>As for the encoding section, you have a valid point. I will adjust it
>for the -01 draft including the ABNF grammar in the appendix.
>Martin Duerst wrote:
>> Hello Yakov,
>> Good work. Two comments:
>> - I think registering just text/cvs should be enough.
>> One point of the registry is to reduce the number of values
>> used for the same thing to one. The registration can serve
>> to converge to that.
>> - I like the realism in the "Encoding considerations" section,
>> but shouldn't you at least mention LF also for line ends,
>> or just leave out specific values?
>> Regards, Martin.
>> At 10:58 05/02/03, Yakov Shafranovich wrote:
>> >Hi all,
>> >I noticed that the MIME registration tree has an entry for
'text/tab-separated-values' but for some reason doesn't have any
registration for 'text/comma-separated-values' or 'text/csv. I have been
reviewing some of the stuff out there and so far have seen the following
>> >I was wondering if there would be any value to formally register two
of the CSV types being used from the text tree (text/csv and
text/comma-separated-values). I put together a small draft which hasn't
posted to the ID repository yet. A copy should be available here:
More information about the Ietf-types