On 9/24/14, 12:16 PM, Matthew Wild wrote:
On 24 September 2014 16:38, Peter Saint-Andre - &yet <[email protected]> wrote:
On 9/24/14, 9:34 AM, Hund, Johannes wrote:
Wow, finally some more „exotic” M2M use cases. Sounds interesting your
project!
In terms of message load I could think tigase might be also be worth a
look – it has downsides on configurability and extensibility from what
I’ve heard though.
I see that certain pattern, and maybe it would be good to have a Matrix
of the brokers with some infos like scalability in terms of user count ,
scalability in terms of message load on the other column, clustering,
ease of configuration, code base maturity, etc.
Or maybe just extend the present on xmpp.org in terms of ‘”primary
design goals”.
Would you think that to be senseful or just me asking for a flamewar?
Mostly just asking for a flamewar. It is difficult to sort truth from fancy
(or marketing claims). It is also difficult to find people who will keep the
information objective and up to date.
Agreed. For example, there is already such a matrix on Wikipedia. It's
certainly based more on marketing claims than the truth. As a result I
stay well away from it (which means Prosody often looks worse than it
is compared to the other implementations, but... principles!).
To maintain an accurate and objective matrix I think the best solution
would involve automated testing (which is also liable to be incorrect
of course, but it could be improved over time - and generally bugs
would affect all server implentations equally). Such an effort would
not be a trivial undertaking, and I don't think it would be worth it.
Probably that's why we have talked about it on and off for years but
never made it happen.
Peter
