Re: [apps-discuss] [Uri-review] tn3270

[Joe Salowey <[e-mail]>]

On Nov 19, 2010, at 10:44 AM, Alfred HÎnes wrote:

> Michael, all,
> 
> nobody (ehem, at least almost nobody) doubts that TN3270 service
> is in use today.  But that isn't the question being discussed in
> this thread.
> 
> The questions of interest are:
> 
> a)  Are 'tn3270' URIs used ?
> 
[Joe] Yes, there are tn3270 terminal emulators that install themselves as a helper application for a web browser similar to a telnet application.  


> b)  Where is the specification and a registration document ?
> 
[Joe] Don't know


>    Or is there only a semi-secret oral tradition that nobody has
>    deemed worth being written up in a URI scheme registration
>    document brought to IANA ?
> 
> and secondary to a), conditional under b) :
> 
> c)  Is it expected that new implementations of software making
>    use of 'tn3270' URIs will be done ?
> 

[Joe] Sure, people are still selling terminal emulators

> d)  Does the IETF want to encourage new URI implementations to support
>    the 'tn3270' URI scheme (if it has been actually defined) ?
> 
> [ And I personally would like to add:
>  If there actually is demand, why is there no specification that would
>  allow an independent, yet interoperable, new implementation?
>  But that's also not the basic question.
> ]
> 
[Joe] Why would they want to discourage it?  there may be no definition perhaps because the URI scheme is very similar to telnet (which was undocumented in the IETF for a long time).  

> Unless all of a) through d) can be answered in an affirmative manner,
> the correct conclusion to draw indeed would be to move the reserved
> scheme name to Historic.
> 
> There seems to be some sublime, unspecific FUD about "Historic".
> "Historic" for an RFC or an IANA registry entry does not mean
> "no implementations" or "not used"; it is an indication that the
> specification or technology
> -  has been or is going to be superseded by something newer
>   (you just have confirmed e.g., that indeed for "System i" there
>   is a preferred, more recent successor technology for TN3270 service,
>   and if 'tn3270' URIs were needed, there might be something other
>   for use with TN5250 as well, isn't it?), and
> -  the IETF does not recommend any more general implementation of
>   the older technology.
> 

[Joe] I guess I don't particularly care if it is moved to historic or not because I don't think it will effect if anyone will use it or not., but I don't think either of the two above apply.  There is tn5250 and some use a URI for this as well, but it is different than 3270 and does not necessarily supersede it.  


> Anybody is free to implement a Historic specification, but they should
> not expect everybody to do so; and everybody is free to [continue to]
> use implementations of a specification that has been moved to Historic
> status.
> (To reinforce it, "technology" here means use of 'tn3270' URIs,
> not use of TN3270 service.)
> 
> I also rely in my work on lots of protocols that now are considered
> Historic; this does not cause a disease; it's simply that neither
> I myself nor my customers have an expectation that these are
> implemented in common new hardware/firmware/software.
> (E.g., RTELNET, RFC 818, comes to mind -- still used with numerous
> factory floor devices -- but Status: HISTORIC.)
> 
> A short note on IANA processing:  Since a while, IANA can manage
> timed processes; this hasn't been so in the past.  In more recent
> IANA registries that make use of this capability, "provisional"
> entries are bound to an expiry date, and without updates of such
> entries or conversion to regular entries, they automagically
> disappear from the registry after the time-out.
> Conceptionally, this apparently was the idea in RFC 1738, where it
> talked about "reserving the name, waiting for future specification"
> (which had been expected to be provided in a reasonable timeframe).
> 
> 
> To repeat my recommendation to stakeholders from my previous posting:
> 
> If there _is_ an agreed-upon spec for 'tn3270' URIs, write it up in a
> short URI scheme registration draft (that should be possible in no
> more than half a dozen of pages) and pursue publication as an RFC
> (I assume that our APP ADs would support this in the same manner they
> offered support for "blindly" moving to Historic a non-specified
> scheme, via a dedicated memo).
> 
> If such effort doesn't happen, this will be taken as a strong
> indication that there is no sustained interest in the provisionally
> registered 'tn3270' URI scheme name and it actually SHOULD be moved
> to Historic.
> 
> 
> Kind regards,
>  Alfred.
> 
> -- 
> 
> +------------------------+--------------------------------------------+
> | TR-Sys Alfred Hoenes   |  Alfred Hoenes   Dipl.-Math., Dipl.-Phys.  |
> | Gerlinger Strasse 12   |  Phone: [phone], Fax: -18         |
> | D-71254  Ditzingen     |  E-Mail:  [e-mail]                     |
> +------------------------+--------------------------------------------+
> 
> _______________________________________________
> apps-discuss mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/apps-discuss