Re: CRC

• Date: Wed, 14 Feb 1996 02:11:11 PST
• From:
  mark@darkhorse.cv.com (Mark R. Bowyer)
• Cc: jft@psychology.nottingham.ac.uk
• Content-MD5: /OoE1TI0WGj4cW1OpMKt8Q==
• Content-Transfer-Encoding: 7bit
• Content-Type: text/plain; charset=us-ascii

> I wasn't planning on using this to make sure the text got to the other MOO
> correctly, I know it will since the network already uses some sort of CRC
> checking.  What I wanted was for BOTH MOOs to generate a CRC of the object's
> code.  If the CRCs match, the MOO will not update anything.  If it doesn't,
> then the host MOO will update the client MOO's code... That way, this group
> of objects will remain exactly the same on both MOOs, incase I decide to
> alter one, I only have to do it once.

Neat idea, I understand what you're after now...

At MOOtiny, as part of our Multi-Owner Object system, each verb has a string 
tacked on the end, no matter how it is edited, that states when it was edited 
and by whom.  This would give some indication of verbs needing updating 
elsewhere, I spose.  But once we get to following changes to props too, then 
definately we'd need something that produced a checksum/CRC for those verbs/ 
props declared on that object to compare.  The declared on object only bit 
should stop problems with the creation time always being different, etc.

But the big problem, especially on objects that are maintained by a group, or in 
the case of most of my objects where the users can expand the object themselves, 
what do you do in the case where more than one version of an object on different 
MOOs have been updated within a checking period?  Would it be possible to 
include some code here for merging differences in props to both?  And what of 
verb code that *has* to be different on different MOOs due to core differences?  
The different ways integration is handled comes to mind...  Can we put in 
machine-readable comments around blocks that need not be included in the 
checksum, and aren't replaced in updates?

Just a thought or too to make this job *loads* more complex.  I've seen people 
discussing this before - an easier way of keeping code on multiple MOOs the same 
over them all - and it's good to see someone actually doing something about it.  
I just wanted to mention at the early design stages some of the problems that 
have occured to me before on this...

Good luck,

                                      !
                                     ,,,
                                    (. .)
/--------------------------------o00-(_)-00o----------------------------------\
|    o o   					Computervision Ltd,	      |
|    /v\ark R. Bowyer.				Innovation Court,	      |
|    `-'	mbowyer@ukhayq.cv.com		New Street,		      |
|    		Tel: +44 (0)1256 301188		Basingstoke, HANTS, RG21 7DP  |
|		Fax: +44 (0)1256 301140		United Kingdom		      |
|GCS/E/TWd-Hs+g-au@a-w--v+C++US+++++$E+po--Y+t+5+tvb++e++u**h----fr+++n++x++++|
|-----------------------------------------------------------------------------|
|    "I may be lost, I may be blind, I may be cursing most of the time, but   |
| that's half the fun.." - `I can't imagine the world without me' - Echobelly |
\-----------------------------------------------------------------------------/
(My Geek Code is FAR shorter - does that make me less of a man?)
(see 
http://krypton.mankato.msus.edu/~hayden/.codes/geek3.1.html)

• Prev: RE: CRC
• Next: Re: CRC
• Index: MOO-cows
• Thread: MOO-cows