On Tue, Apr 7, 2009 at 8:09 PM, Michael Markstaller <[email protected]> wrote:
> Paul, I'll have a look on this.
>
> I guess in contrast to C-programming I have some understanding of SNMP,
> working with it for many years now. Its without question something strange
> ;) But IMHO also, once understood very practical, scalable and good; owfs is
> rather predestinated for it getting and settings data from/to sensors..
>
> Although it's from a copy'n'paste scripting perspective guy like I am no
> real big issue to realize this working but very inefficient there is IMHO a
> much better way. Net-snmp is very widely used today (at least in the scope
> of snmp-addictives)
Net::SNMP would be a fine approach to this program. Coupling it with owperl
or OWNet would be a fast way to accomplish snmp support. In fact, we could
later decide if a C implementation is even needed. The perl program could be
considered a prototyping step.
>
> Let me put together a reasonable presentation of the owfs-data during the
> next days. I could also take care of the MIB-part if needed.
I'd like to use the same naming scheme as the rest of OWFS, as much as
possible. That's part of the power of OWFS.
> If this will get anything further it might make sense to register an
> enterprise number for owfs, so that this stuff could be located somewhere
> meaningful, which is IMHO the place where it should be: (
> http://pen.iana.org/pen/PenApplication.page) it's AFAIK still free.
>
Submitted.
> Although I could live with a "limited hack to just do it", I'd be happy to
> contribute as far as possible.
> While playing around further today I already ran into the first problem
> with floats in the external scripting only being possible by some obscure
> opaque_xxx - or as integer divided by 10/100/1000 (which was and is
> braindead, I always wondered until the readings today, why most of the
> things I know present a temperature as int=float*100 or as string..)
We use floats for temperatures, humidity, voltages, current, and pressure.
>
> But as far as I can see the C-API and docs for (agentX-)Plugins is quite
> comprehensive and well documented now.
>
> The presentation is in this special case where we have not only a few
> values most likely the biggest thing to consider, owfs is very predestinated
> and structured for that already. But with a general approach there are a few
> more things to be considered like I/Os (what I haven't), I'll think about
> how to present this, I guess I've seen at least enough bad implementations
> to say this should take more than tonight ;)
>
> regards Michael
>
> > -----Original Message-----
> > From: Paul Alfille [mailto:[email protected]]
> > Sent: Wednesday, April 08, 2009 12:38 AM
> > To: [email protected]
> > Subject: Re: [Owfs-developers] OWFS & SNMP
> >
> > I can program in C, but SNMP seems rather forbidding.
> > Specifically, how the readings would be presented.
> >
> > I found this example from ora-ltd:
> > http://www.ora-ltd.com/microweb/mib.html (The MIB is
> > http://www.ora-ltd.com/microweb/MTECH-DSTEMP-MIB.mib)
> >
> > I assume we would used the owfs names (i.e. temperature,
> > type, id, address) for the lowest level data.
> >
> > So can you explain more of how this should work?
> >
> > Paul Alfille
>
------------------------------------------------------------------------------
This SF.net email is sponsored by:
High Quality Requirements in a Collaborative Environment.
Download a free trial of Rational Requirements Composer Now!
http://p.sf.net/sfu/www-ibm-com
_______________________________________________
Owfs-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/owfs-developers
