About http://osmocom.org/issues/1757#note-4
With the current naming scheme, the logical name for a fully dynamic TCH/F-TCH/H-PDCH channel would be
GSM_PCHAN_TCH_F_TCH_H_PDCH
timeslot 2 phys_chan_config TCH/F_TCH/H_PDCH
Is it asking for trouble to just name it DYN instead?
GSM_PCHAN_DYN
timeslot 2 phys_chan_config DYN
I'm in (at least) two minds about it, opinions welcome...
~Neels
On Wed, Jul 13, 2016 at 1:18 PM, Neels Hofmeyr nhofmeyr@sysmocom.de wrote:
About http://osmocom.org/issues/1757#note-4
With the current naming scheme, the logical name for a fully dynamic TCH/F-TCH/H-PDCH channel would be
GSM_PCHAN_TCH_F_TCH_H_PDCH
timeslot 2 phys_chan_config TCH/F_TCH/H_PDCH
I personally prefer this option, because it doesn't leave a place for doubt. And ti's not something you would type 100x a day, so making it short is not a real optimization.
Is it asking for trouble to just name it DYN instead?
GSM_PCHAN_DYN
timeslot 2 phys_chan_config DYN
Reading this, I would be asking myself "DYN - what is that? Dynamic? Ok, Dynamic what? Channels? Ok, which channels? Can it have SDCCH in there? Hmm.. Don't know. Let look into the docs. Docs say no. Can it be SDCCH in the next release of the software? Huh, don't know, I better to put something more specific here, so I'm sure what's happening. But how to do that?"
Users who don't ask those questions will ask a lot of other questions like "I thought DYN is doing XYZ, but it not! Help!"
Hi Alexander and others,
On Wed, Jul 13, 2016 at 04:59:11PM -0400, Alexander Chemeris wrote:
phys_chan_config TCH/F_TCH/H_PDCHI personally prefer this option, because it doesn't leave a place for doubt. And ti's not something you would type 100x a day, so making it short is not a real optimization.
I also think it should be kept this way for consistency with other pchan names. From a VTY syntax POV, the '/' are ugly in the name, and we never use that anywhre else in the VTY but the channel config - but the '/' is part of the official specified names, so I chose to use them with the existing channel types when originally writing that code.
Reading this, I would be asking myself "DYN - what is that? Dynamic? Ok, Dynamic what? Channels? Ok, which channels? Can it have SDCCH in there? Hmm.. Don't know. Let look into the docs. Docs say no. Can it be SDCCH in the next release of the software? Huh, don't know, I better to put something more specific here, so I'm sure what's happening. But how to do that?"
The VTY offers a help string at tab-completion which could have the anwer to this, so I don't think that point is such an issue.
On Thu, Jul 14, 2016 at 02:40:10AM +0200, Harald Welte wrote:
Hi Alexander and others,
On Wed, Jul 13, 2016 at 04:59:11PM -0400, Alexander Chemeris wrote:
phys_chan_config TCH/F_TCH/H_PDCHI personally prefer this option, because it doesn't leave a place for doubt.
I also think it should be kept this way for consistency with other pchan names.
Agreed. I thought, maybe we want to add more pchan types so having a shorter DYN name could be useful. But having a consistent, explicit, logical name is better.
From a VTY syntax POV, the '/' are ugly in the name
Yeah, especially since it looks like "TCH or F_TCH or H_PDCH" :) But one gets used to it quickly.
~Neels
-
Alexander Chemeris -
Harald Welte -
Neels Hofmeyr