Jody Garnett ha scritto:
> Andrea Aime wrote:

> Don't follow you about generating code? The Transaction format is quite 
> clear though - broken down into delete, update, add ... think this one 
> is perfect. If you want to talk further I suggest we make a real example 
> and work through both approaches; and then probably write an example 
> parser (are you targeting a Java application programmer or AJAX here?)

What I meant, the clients does not have parsing code for Transaction 
already in place, just the code that makes Transaction requests.
If we call this GetDiff response format "transation", I'm only saying
that we have either to provide another one that's easier to deal
with when you want to _show_ the differences, because otherwise we
force clients to parse that stuff and apply it to features for revision
M in order to show those at revision N... basically we're making the 
client do the same work as a server, no?

>>> This is consistent with Galdos cascading WFS-T approach and would be 
>>> a *great* benefit for keeping servers in sync.
>>> (Please consider this idea).
>> I don't follow you... why Galdos matters in a GetDiff request?
> Because they have covered this territory before us; and thus we can say 
> the idea is implementable - and better yet deployed.

Sob... Jody, you are assuming that I know Galdos products, but they site
is full of marketing fluff an no technical stuff, so I simply turned it
down as useless. From your comments, I gather they have versioning 
already? Or what?

>>> Rollback - don't do it, use GetDiff
>> I disagree here for two reasons:
>> * doing a rollback in the server is very quick, it's just a matter of
>>   a couple of queries (much faster, the more the bigger you revert is)
> Ah you putting your performance concerns before the design; I had hoped 
> not to introduce any more concepts for users to concern themselves with. 
> Having operations that slot together etc...

My point of view was the opposite. IMHO, not having transaction may save 
work on the client side, but makes life harder for the client.
Rollback it's already in the vocabulary of users, and when you go and 
try to do it you've always have to deal with some other concept (dance
of doom in cvs, reverse merge in svn).

>> * if the rollback is big, the transaction request to do the actual
>>   work can be a huge gml document.
> Like this point; at least *define* the service in terms of being exactly 
> the same as getting the difference and applying it using transaction 
> (makes the description very clear).

Hum... I think I like the concept but I can't imagine what it would look
like. I was thinking, having the equivalent of svn merge (directly 
opearting on the server) would not be bad... so, would a new Transaction 
element designed to do merges work?

Cheers
Andrea

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Geotools-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geotools-devel

Reply via email to