Text 3823, 44 rader
Skriven 2009-11-17 23:33:40 av mark lewis (1:3634/12.0)
   Kommentar till text 3818 av Kevin Gates (1:261/38)
Ärende: FTS-4009
================
>> It is clear that some of the examples of "valid" usage are not valid.
>> It may be more accurate to say that they are examples of what might be
>> found in the wild at the time the document was written.
> Maybe it is better to fix the standard so that it contains the
> examples that are valid now?

 KG> I agree completely.  Since the proposed standard contained valid
 KG> examples, I'm not sure why these bad examples were included in the
 KG> actual standard.  I suspect someone made a mistake.

no... not really... examples of live in the wild VIA lines were given BECAUSE
they are/were live and in the wild... remember, the VIA lines were never
intended to be parsed by any automation tools... their existance is/was for the
sysops operating the machines thru which the message traveled...

>> While I personally feel the 10 character limit on these fields is
>> arbitrary and silly, changing it now could potentially break other
>> software that was written "to spec".
> How can it break the software? If the software sticks to the present
> standard then it "may need to parse" at least the examples found in 
> the standard and declared to be valid. And the examples break the 10 
> character limit.

i can possible that the "mistake" you mention above is placing a limitation on
the "field lengths"... but i must also point out, again, that until this
document came along, there was no template or format specified... as long as
the line started with the proper ^AVIA format, contained the name of the
program attaching the VIA line and a timestamp of some form... 

 KG> Making assumptions about how other programmers code their products
 KG> is dangerous. How can you be sure that the change you proposed
 KG> would have no impact on any existing software?  The safest course
 KG> of action is to just fix the examples and leave the existing Via
 KG> line field specifications as they are. 

actually, the safest thing is for programmers to NOT parse the lines in the
first place since they were never designed for such...

remember that old adage... be loose with what you accept but tight with what
you send...

)\/(ark

 * Origin:  (1:3634/12)