Re: [Uri-review] 'ftp' and 'file' URI schemes still needs specifying

[Graham Klyne <[e-mail]>]

I distantly recall that the working group that led to RFC 3986 spun out some
activities, or at least proposed such, to try to regularize both ftp: and file:.

I have a recollection  that Paul Hoffman edited a round or two of Internet
Drafts, but the effort ran out of steam.  I also recall a suggestion that rather
than trying to standardize file:, a useful step would be to describe the
varieties of existing behaviours as an informational document, but again there
was no-one with the time and/or energy to push this forward.

To give a sense of timeframe for these discussions, the most recent message
about specific activities (or not) that I can find is:

   http://lists.w3.org/Archives/Public/uri/2004Feb/0014.html

In response to this message (from yours truly) that gives some sense of the 
discussion around file: URIs:

   http://lists.w3.org/Archives/Public/uri/2004Feb/0011.html

After that, I'm not seeing much directed discussion.

#g
--


Mykyta Yevstifeyev wrote:
> 15.05.2011 11:20, "Martin J. Dürst" wrote:
>> Hello Mykyta,
>>
>> Providing specifications for ftp: and file: would indeed be a good 
>> thing. But please be aware that besides the explicit publications you 
>> found, there have also been other efforts to specify these schemes, 
>> and they never got very far. Of the two, I'd guess ftp: is easier, but 
>> I might be wrong. I definitely know that the file: scheme will be 
>> tough; there is quite some variation among browsers, not to mention 
>> other software, and it will be difficult to provide the right balance 
>> between prescriptive and descriptive approach.
> Thanks for response, Martin.  I recall some discussions on [e-mail] 
> list regarding the 'file' URI scheme and can say it is really hard to 
> specify it; that person who will undertake this work should probably 
> spend a large amount of time on investigation of the issue and putting 
> this on paper.  Maybe the difficulty is a great number of variants 
> existing for file URIs, ie. absence of unified syntax, first of all.  
> The situation is radically different with ftp; there is one well-known 
> and widely-accepted syntax of it.  After we'll have defined these two 
> schemes, it will make sense to change RFC 1738's status to Historic, 
> considering anything of its content will be "shared" between other 
> documents, but file URI scheme will be stumbling block here.
> 
> Mykyta Yevstifeyev
>>
>> Regards,   Martin.
>>
>> On 2011/05/15 14:14, Mykyta Yevstifeyev wrote:
>>> Hello all,
>>>
>>> The first mention of 'ftp' URI scheme in the IETF document is probably
>>> RFC 1738 (http://tools.ietf.org/html/rfc1738), if I'm right. As far as
>>> you can see, this RFC is formally obsoleted by RFC 4248 and RFC 4266,
>>> but in fact RFC 2396 et seq., meaning RFC 3986 (formally they updated
>>> RFC 1738), replaced it. Among other, RFC 1738 document specified several
>>> URI schemes, such as widely-used 'http', 'mailto' etc. This table
>>> summarizes the state of schemes specified by RFC 1738. ("Specification"
>>> is the most current RFC defining some scheme).
>>>
>>> /Scheme name
>>> / /Protocol
>>> / /Specification
>>> /
>>> http
>>> HTTP (RFC2616)
>>> RFC 2616
>>> gopher
>>> Gopher (RFC1436)
>>> RFC 4266
>>> mailto
>>> N/A
>>> RFC 6068
>>> news
>>> NNTP (RFC 3977)
>>> RFC 5538
>>> nntp
>>> NNTP (RFC 3977)
>>> RFC 5538
>>> telnet
>>> Telnet (RFC 854)
>>> RFC 4248
>>> wais
>>> WAIS (RFC 1625)
>>> RFC 4156
>>> file
>>> N/A
>>> *RFC 1738*
>>> prospero
>>> Propsero (non-IETF)
>>> RFC 4157
>>> ftp
>>> FTP (RFC 959)
>>> *RFC 1738
>>> *
>>>
>>> :
>>> You may see, there are two schemes in this list specified by formally
>>> obsoleted RFC 1738 (even though it is actually PS). They're probably the
>>> only two schemes (not considering 'afs', but it's in provisional
>>> regsitry) that are listed at IANA registry, in the Permanent category,
>>> with a reference to obsoleted RFC (maybe, 'fax' can be considered to be
>>> so as well; but is it de facto deprecated and hsitorical). 'ftp' and
>>> 'file' schemes are quite widely-used; obsolete RFC 1738 is an actual
>>> specification for them.
>>>
>>> There has been an effort to specify this schemes in separate docs.
>>> (https://datatracker.ietf.org/doc/draft-hoffman-ftp-uri/,
>>> https://datatracker.ietf.org/doc/draft-hoffman-file-uri/), but they
>>> resulted in nothing (unlike eg. RFC 4248, as a part of the same effort,
>>> if I'm right). Considering this, should an attempt to provide these
>>> schemes an up-to-date specification be undertaken?
>>>
>>> Mykyta Yevstifeyev
>>>
>>>
>>>
>>> _______________________________________________
>>> Uri-review mailing list
>>> [e-mail]
>>> https://www.ietf.org/mailman/listinfo/uri-review
>>
> 
> _______________________________________________
> Uri-review mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/uri-review
>