We use distributed files and yes a master file is created. The DEFINE.DF
command creates a type 27 file and a dictionary file is created that is
a type 30 file.
As a matter of fact we use the one dictionary for all of the part files.

-----Original Message-----
From: Anthony Youngman [mailto:[EMAIL PROTECTED] 
Sent: Friday, June 22, 2007 6:02 AM
To: [email protected]
Subject: RE: [U2] [UV] Trigger(s) on distributed file

Odder and odder.

I don't ever remember seeing a file at OS level with the name of a
distributed master file. PI *or* UV. As implemented by Prime, iirc, any
partfile could be used as a master to get the algorithm definition etc,
and you were warned that a VERY QUICK way to corrupt your distributed
file was to use FORCE with DESIGN.DF unless you knew exactly what you
were doing - precisely because a part file could belong to multiple
masters and if the masters weren't in sync the result was a terrible
mess...

Cheers,
Wol

-----Original Message-----
From: phil walker [mailto:[EMAIL PROTECTED] 
Sent: 22 June 2007 11:20
To: [email protected]
Subject: RE: [U2] [UV] Trigger(s) on distributed file

Sorry Wol,

Yes, I did think you meant the Pick style multi-part, sorry about that.

But as far as the distributed files as managed by DF they have always
had a master file of type (27 or 41 - cannot remember which ) which was
basically just a header and I think had the algorithms etc in them.

Dave Meeks or Glenn Herbert could probably confirm if they are still
around. Palmer Chapelle was the engineer who had most input in
implementing them in UV from PI.

Cheers,

Phil

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Anthony
Youngman
Sent: Friday, 22 June 2007 10:16 p.m.
To: [email protected]
Subject: RE: [U2] [UV] Trigger(s) on distributed file

Then things really are confusing ... I'm talking about the file-type
introduced by Prime, and officially managed using the DEFINE.DF command.

By "multi-part", do you mean the Pick style NEW.ACCOUNT,PIOPEN type
file? I'm certainly not thinking of those - they were never in PI and
I've never had cause to use them. I wouldn't presume to know anything
about them.

Distributed files, as known on Prime, never to my knowledge had any
physical existence other than as individual part-files. As far as I was
aware, the UV implementation had similar characteristics ...

Cheers,
Wol
-------
u2-users mailing list
[email protected]
To unsubscribe please visit http://listserver.u2ug.org/
-------
u2-users mailing list
[email protected]
To unsubscribe please visit http://listserver.u2ug.org/

Reply via email to