On 11/28/06, Gerrit Renker <gerrit@xxxxxxxxxxxxxx> wrote:
Ian McDonald wrote:
| Why is it broken? The RFC says that the user can have the option of
| setting the packet size s, so it is not experimental. Yes it is only
| for CCID3 but there are already lots of things that are CCID2 or CCID3
| only.
... and that is the problem I am referring to. From the local perspective of adding
this or that socket option, nothing is probably too much of a burden. But there
is no documentation for these socket options (I have at begun with documenting some
sysctls). Hence one can not, with a good conscience, argue that the programming interface
is `as easy as UDP'. I think that should be the guideline and that is what I am proposing,
leave everything out which is not strictly and essentially requrired. Otherwise, it will
confuse users.
I had meant for a long time to put up my updated ttcp clone but I will put nothing online
until the socket API is not ironed out and documented.
And there are some credits to be had for people daring enough to document the socket API -
the DCCP user guide is apparently well well out of date.
In short: my suggestion is to keep an experimental patch for this and I would even offer to
keep one up-to-date and online, if in return we can simplify the socket API. Does
this sound like a more convincing argument?
Fair enough, I think we should go this way for now, please post here
the patch that provides the experimental feature of explicitely
setting the packet size, interested people can try and use it and
report their findings, later we can get back and possibly merge the
patch if it proves useful.
- Arnaldo
-
To unsubscribe from this list: send the line "unsubscribe dccp" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
