🚀 go-pugleaf

In article <V7aVrAYOBh107h@aescon.com>, aesmoot@aescon.com (Art Smoot) writes:
> In <19930113.113849.111@almaden.ibm.com> wohler@vnet.ibm.com (Wayne L. Wohler) writes:
>>In <9301131516.AA18603@mingus.techno.com> "Steven R. Newcomb" <srn@techno.com> writes:
>>>I disagree.
> 
>>I agree with Steve.
> 
> Clearly there is some disagreement on this thread about whether any
> SGML editor should always store ASCII SGML or an "internal format" and
> then export ASCII SGML as needed.  Whichever format is used for storing
> the SGML, IMHO, depends on how the overall document management system
> handles the data.  
Agreed...

> If the SGML editor uses an internal format which can
> be stored, then then it is incumbant on the "system" to guarantee that
> the SGML file always be synchronized with the internal format file.

Why bother...?

Why not do all updating, editing, additions, etc on the SGML editor
with its own internal "object" format. When there is need for an ASCII SGML
file (eg. to send outside the organisation, to use in a package
incompatible with the internal format, etc), then just "output"
the ASCII-based marked-up file(s). This should be only done on a
similar basis to outputting a PostScript representation from a 
work processor.

> For any complex document created in a multi-author environment this is
> essential.  Imagine one author updating the ASCII SGML file and another
> updating the same document's internal format file.  Never happen?
It *should* never.

Responsibility for updates has to end somewhere.
eg. Are you going to consider a change to a PostScript file
as a valid change to a source ?

---
Liam Relihan,           | Voice:  +353-61-333644 ext. 5015 |
C.S.I.S.,               | Fax:    +353-61-330876           |    W H I T E
University Of Limerick, | E-mail: relihanl@ul.ie           |    S P A C E
Ireland.                |                                  |