[ome-devel] Shoola-back-end consistency issues

[Ilya Goldberg]

On Jul 7, 2005, at 9:50 AM, Josh Moore wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Harry Hochheiser wrote:
>
>> fair enough, but I want to be very clear on this: implementing refresh
>> buttons in the various agents is neither simple nor fun. Given n 
>> agents
>> with different DTOs and data requirements, and a stateless server, it
>> will be somewhere between tricky and virtually impossible to handle
>> refreshes in a generalized manner.  Given the data requirements for 
>> some
>> of the Shoola clients, doing a "dumb" stateless refresh will be
>> incredibly slow and wasteful. If I've downloaded a 5MB dataset, and 
>> one
>> image gets added, I simply can't say "re-request the whole dataset".
>> The client performance would go to hell, and we'd spend lots of
>> bandwidth needlessly shuffling things back and forth.  I might be able
>> to solve this special-case for certain cases, but generalizing it 
>> would
>> be difficult, if it's even possible.
>
> Harry, I would punt in the other direction to be honest. I'd implement
> the fastest (in terms of development) "refresh" button you can, and
> we'll generalize when the server architecture is there. From the user
> perspective, restarting Shoola to see the new images in a Dataset is a
> LOT slower than pulling the whole dataset down the wire.

That's very true, and what I had in mind.  Its the lesser of two evils. 
  We can't afford to punt on consistency.
-I


>
>  -J.
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.5 (GNU/Linux)
> Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
>
> iD8DBQFCzTMhIwpkR5bKmAsRAk6mAKCgT8qd4obS+RAMK7BZcEOaQSPBCgCeMUp3
> B2CaeTkzeG6OZlVdZC6sLgc=
> =uvQv
> -----END PGP SIGNATURE-----
> _______________________________________________
> ome-devel mailing list
> ome-devel at lists.openmicroscopy.org.uk
> http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-devel
>