Number: 2257

Date: 25-Sep-84 12':21':40

Submitter: Sannella.PA

Source: vanMelle.pa

Subject: Should sysout CREATIONDATE be changed when new ucode is spliced in?

Assigned To: 

Attn: vanMelle, Charnley, Masinter

Status: Open

In/By: 

Problem Type: Design - Impl

Impact: Annoying

Difficulty: 

Frequency: 

Priority: Perhaps

System: Operating System

Subsystem: Generic File Operations

Machine: 1132

Disk: 

Lisp Version: 24-Sep-84 09':49':01

Source Files: 

Microcode Version: 5124

Memory Size: 4096

File Server: 

Server Software Version: 

Disposition: 

Description: '
Date': 24 Sep 84 15':12 PDT'
From': vanMelle.pa'
Subject': Re': New DLion uCode'
In-reply-to': charnley.pa''s message of 17 Sep 84 11':54':16 PDT (Monday)'
To': charnley.pa'
cc': vanMelle.pa, LispSupport'
'
But there''s a way to tell the server not to change the creation date.'
'
So I put a new version of SPLICE on <LispCore>Sources> with this modification in it.  Using this new version should help reduce confusion about when a sysout was actually made (I think that it will still update the sysout''s Write Date, so you can tell someone touched it, but the Creation Date will remain that of when the sysout was originally made).'
'
	Bill'
'
-----'
'
Date': 24 Sep 84 16':54 PDT'
From': masinter.pa'
Subject': Re': New DLion uCode'
In-reply-to': vanMelle.pa''s message of 24 Sep 84 15':12 PDT'
To': vanMelle.pa'
cc': charnley.pa, LispSupport.pa'
'
I *liked* it that splicing in new microcode changed the creation date.... it isn''t a null-content change, after all.'
'
E.g., "do I have to update my DLion?" depends on whether the latest sysout has a different creation date than the one I already have..'
'
-----'
'
Date': 24 Sep 84 17':00 PDT'
From': vanMelle.pa'
Subject': Re': New DLion uCode'
In-reply-to': masinter.pa''s message of 24 Sep 84 16':54 PDT'
To': masinter.pa'
cc': vanMelle.pa, charnley.pa, LispSupport.pa'
'
For this, I offer you the write date.'
'
Meanwhile, Dolphin/Dorado users look at the creation date and think there IS a new sysout to load, when there really isn''t.'
'
I''m not really sure what the right thing to do is (after all, it is true that the content of the file changed, at least for some users of the file).'
'
	Bill'
'
-----'
'
Date': 24 Sep 84 21':50 PDT'
From': JonL.pa'
Subject': Re': New DLion uCode'
In-reply-to': vanMelle.pa''s message of 24 Sep 84 17':00 PDT'
To': vanMelle.pa'
cc': masinter.pa, charnley.pa, LispSupport.pa'
'
I''m with you here, Bill.  Splicing is precisely the kind of operation that necessitated a distinction between the "CreationDate" and the "WriteDate" file properties.'
'
-- JonL --'
'
-----'
'
From': Masinter.pa'
Date': 24-Sep-84 22':38':10 PDT'
Subject': Re': New DLion uCode'
In-reply-to': JonL''s message of 24 Sep 84 21':50 PDT'
To': JonL'
cc': vanMelle, masinter, charnley, LispSupport'
'
When I install a sysout on my DLion, it records the creation date as it existed on the file server, and the "write date" as the date it got brought over and installed on the DLion. In general, in the distributed environment, "creation date" refers to the last time content changed, while "write date" refers to the date when the contents got recorded on the particular medium. This is a reinterpretation of "write date" from what it used to mean in the time-sharing environment when things didn''t move from one place to another.'
'
I can understand how Dolphin/Dorado users object to having new microcode supposedly affecting content when it doesn''t for them, but I think that its a pretty weak case.'
'
We made it so that rewriting the file map doesn''t change the creationdate, and that seems this side of the edge, since one could argue that the filemap is completely derived information. I don''t think, for example, that running HITATOMS should leave creationdate unchanged, and splicing new microcode seems like it really does change content.'
'
Did he splice it or did he not? When did it get spliced? Is this a version before or after splicing?'
'


Workaround: 

Test Case: 

Edit-By: 

Edit-Date: