[rfc-i] extension for xml2rfc files?
pkyzivat at alum.mit.edu
Wed May 7 08:11:47 PDT 2014
On 5/7/14 10:54 AM, Julian Reschke wrote:
> On 2014-05-07 16:18, Paul Kyzivat wrote:
>> On 5/7/14 2:12 AM, Julian Reschke wrote:
>>> On 2014-05-07 02:03, Nico Williams wrote:
>>>> On Tue, May 6, 2014 at 3:17 PM, Paul Kyzivat <pkyzivat at alum.mit.edu>
>>>>> [...]. IMO this is inappropriate - the extension ought to be more
>>>>> specific to the
>>>>> expected format of the document. [...]
>>>> .xml _is_ specific to the format of the document: XML. You meant
>>>> schema, I know :) but still: the schema/DTD is declared in the XML,
>>> It's not in general, and v3 doesn't have an official DTD.
>>>> therefore it can be "tasted", so why pollute the file extension
>>>> namespace? I've never felt a need for this.
>>> I'd like first to understand what problem we are trying to solve by
>>> having a more specific file extension.
>> The same problem that is addressed by having well known extensions for
>> other file types - to help people keep them organized and understand
>> what they have.
> What practical problem is solved by having a different extension?
Are you asking what is solved by having unique extensions for each form
of the document? Or what is solved by changing it from .xml to something
Having a unique extension for each form makes it more convenient to save
different forms of the same document in a single directory and keep them
Having those extension names consistent across authors helps when
documents are passed around.
Continuing to use .xml satisfies those needs. It becomes a (admittedly
small) problem when files representing documents are mixed with files
used for other purposes.
Having a unique extension allows configuring a default action for these
files when opened - possibly choosing a different tool than might be
used for arbitrary xml files.
I'm not claiming this is a major issue. If I am the only one who thinks
this would be helpful then I'll forget it.
> (As we now define a media type a suggested suffix of course could be
> added, I just would like to understand whether this is something that
> really addresses a practical problem).
>> Why do the current canonical form files all end in .txt?
> Why not? They are plain text files, after all.
xml files are also plain text. Why don't we use .txt for them too?
C and Java files are plain text, but we don't normally use .txt for them
More information about the rfc-interest