xAP Talk Transcript 2006 08 29

From XAP Automation
Jump to: navigation, search


29/08/2006 21:02:22 * kevint joins My room
29/08/2006 21:03:08 ‹kevint› mmm hello
29/08/2006 21:03:34 ‹kevint› the data entry window is a but thin here
29/08/2006 21:04:11 * Gregg joins My room
29/08/2006 21:05:00 ‹kevint› ah that made it bigger.
29/08/2006 21:05:06 * kevint joins My room
29/08/2006 21:07:49 ‹James› hi
29/08/2006 21:08:11 ‹James› back from your hols then
29/08/2006 21:08:22 ‹James› have a good time?
29/08/2006 21:08:48 ‹kevint› Yes good thanks, apart from the frisking experience
29/08/2006 21:09:14 * Edward joins My room
29/08/2006 21:09:16 ‹James› hopefully you avoided the 'full' search
29/08/2006 21:09:30 ‹kevint› thank you yes ;)
29/08/2006 21:09:45 ‹James› hi gregg, edward
29/08/2006 21:09:56 ‹Edward› Hi
29/08/2006 21:10:00 ‹kevint› hi all
29/08/2006 21:10:01 ‹Gregg› Hi James and rest
29/08/2006 21:10:46 ‹Edward› My talk window has strange grey blocks on it
29/08/2006 21:10:51 ‹kevint› Thers seems to have been an upsurge in trafficon the group while I have been away
29/08/2006 21:11:08 ‹kevint› still catching up
29/08/2006 21:11:31 ‹James› today was quite a busy day, got home expecting the odd mail and there were loads
29/08/2006 21:11:44 ‹Gregg› <hiding/>
29/08/2006 21:11:49 ‹Edward› not looked yet
29/08/2006 21:11:50 ‹James› E, is tha ie or FF
29/08/2006 21:11:59 ‹Edward› IE
29/08/2006 21:12:30 ‹James› try a shift f5
29/08/2006 21:12:36 ‹Edward› Scrolling off the top now - minor
29/08/2006 21:12:37 * KevinH joins My room
29/08/2006 21:13:17 ‹James› hi k
29/08/2006 21:13:23 ‹kevint› well I certainly support the general concept of a schema which better describes real analog(ue) values
29/08/2006 21:13:28 * kevint joins My room
29/08/2006 21:13:53 ‹KevinH› Hi all,
29/08/2006 21:14:02 ‹KevinH› I thought you'd like that KT
29/08/2006 21:14:13 ‹KevinH› How were your hols BTW ?
29/08/2006 21:14:30 ‹James› from what i read so far, i don't like the data being sent as -10.4C i would prefer teh data and the units to be seperate.
29/08/2006 21:14:34 ‹kevint› good, no computers.......
29/08/2006 21:14:50 ‹James› and maybe have an additional displattext type field
29/08/2006 21:15:30 ‹kevint› I would agree with that and I also generally dislike numeric data being presented in a text= value pair
29/08/2006 21:16:12 ‹James› but in a displaytext= field which is additional to the unit= and value= that would be ideal
29/08/2006 21:16:49 ‹James› and also i would like to see any new schema stay with the concepts of BSC
29/08/2006 21:17:08 ‹kevint› I have said this before but to properly descibe an analog value you really should bhave an associated max and min range value
29/08/2006 21:17:19 ‹James› a .query a /cmd to control a .status for no change and a .event to signify change
29/08/2006 21:17:24 * Edward2 joins My room
29/08/2006 21:17:40 ‹kevint› otherwise how could you present data as a bar graph for instance
29/08/2006 21:17:59 ‹KevinH› One of the effects of BSC is that it discourages people from developing richer schema
29/08/2006 21:18:14 ‹James› yep, very true
29/08/2006 21:19:05 ‹KevinH› but it would also be better to have a range of standard template type schema that might be adopted as universally as BSC
29/08/2006 21:19:20 ‹Edward2› BSC is well established and pretty well understood - seems sensible to build on that by defining some extensions
29/08/2006 21:19:23 ‹James› we could have a near carbon copy DSC schema data status and control
29/08/2006 21:19:25 ‹kevint› There is a danger that BSC will be (mis)used by extenson to descibe data which is not basic status or control as well
29/08/2006 21:20:22 ‹Edward2› So best to define 'legal' extensions before it takes on a life of its own
29/08/2006 21:20:43 ‹KevinH› II think 'sensor' data is pretty basic in concept and if we coudl find a way of extending BSC by adding parameters but not breaking what we already had it would be good
29/08/2006 21:21:17 ‹KevinH› but some are fundamental Level=-10% would break a lot of things
29/08/2006 21:21:52 ‹KevinH› but not the C-Bus gateway ;-)
29/08/2006 21:22:11 ‹Edward2› You are part of the problem!
29/08/2006 21:22:24 ‹James› my vote is a new schema not a bsc extension
29/08/2006 21:22:25 ‹KevinH› many people say that !
29/08/2006 21:22:46 ‹kevint› I have always thought that one of the BSC schemas strengths was the BSC specification itself
29/08/2006 21:23:47 ‹kevint› ie it specifies what is possible pretty clearly
29/08/2006 21:24:40 ‹KevinH› It still ahs some ambiguities but it is fairly thorough
29/08/2006 21:24:45 ‹kevint› Hey edward, are you have ing two votes?
29/08/2006 21:25:49 ‹Edward2› Too many computers to use at once! Got a window on each.
29/08/2006 21:26:18 ‹Edward› He's an imposter - take no notice!
29/08/2006 21:26:20 ‹kevint› perhaps you could lend one to SB...
29/08/2006 21:26:39 ‹KevinH› I rang him again today - no answer :-(
29/08/2006 21:26:52 ‹Edward2› I would if I thought tha't make a difference
29/08/2006 21:27:18 ‹Edward2› Maybe he's got a girlfriend
29/08/2006 21:27:36 ‹kevint› more likely he has caller id ;)
29/08/2006 21:27:50 ‹KevinH› Well - why not give it a try E .. we must get this sorted as it's hampering the whole next step regarding UID's etc
29/08/2006 21:28:11 ‹KevinH› I have runga couple of times with witheld CID too
29/08/2006 21:28:24 ‹Edward2› OK - I'll offer. I miss him too
29/08/2006 21:28:53 ‹kevint› I have equpment available too if required
29/08/2006 21:29:04 ‹KevinH› Ideally if we could get him back - even if just for a short while in maintenance mode - if not then getting that source code..
29/08/2006 21:29:58 ‹kevint› Did I see someone asking about a commandl line app
29/08/2006 21:30:18 ‹kevint› I have posted one that does BSC commands
29/08/2006 21:30:29 ‹Edward2› It's no problem for source code I have (eg, the framework - I have a modified version for .Net 2.0) but it's this stuff I don't have that's the problem.
29/08/2006 21:30:40 ‹KevinH› Just a quick Q James ... as haven't got HS running - do you send heartbeats from each phantom application - I've got the same issue with HV
29/08/2006 21:30:47 ‹Edward2› Not that it'd take too long to write a new hub
29/08/2006 21:31:02 ‹Edward2› But viewer would be a much bigger project
29/08/2006 21:31:54 ‹kevint› KH, from memory of the HS17 connector, each one sends a HB
29/08/2006 21:32:06 ‹James› K, yep i do but i send the main hbeat at 60s intervals and the sub hbeats at 600
29/08/2006 21:32:07 ‹KevinH› E - you have SB's number ??
29/08/2006 21:32:40 ‹kevint› y#u can get it from whois...
29/08/2006 21:32:41 ‹KevinH› and the 'main' app is just one that uses HomeSeer schema ?
29/08/2006 21:32:44 ‹Edward2› Nope
29/08/2006 21:32:55 ‹James› yep
29/08/2006 21:33:09 ‹James› the .master one
29/08/2006 21:34:17 * kevint joins My room
29/08/2006 21:35:04 ‹KevinH› sent by email E
29/08/2006 21:35:12 ‹Edward2› thanks
29/08/2006 21:36:19 ‹Edward2› Plenty of progress on the SlimServer xAP plugin here
29/08/2006 21:36:38 ‹James› excelent - willing tester here...
29/08/2006 21:38:02 ‹KevinH› Ahh yes - reminds me I did look for that IR blaster - no luck so far but will keep looking
29/08/2006 21:38:11 ‹Edward2› Cool - will send when it's stable (just fell over)
29/08/2006 21:38:36 ‹KevinH› Are you still 'hooked in' or a plugin ..
29/08/2006 21:38:40 ‹James› E, does it do display messages, in particular queueing of them?
29/08/2006 21:38:50 ‹Edward2› Not looked at the IR stuff yet but most of the rest has been pretty easy
29/08/2006 21:39:01 ‹Edward2› Queuing?
29/08/2006 21:39:21 ‹James› so it displayes message 1, then 2 then ...
29/08/2006 21:39:40 ‹Edward2› Is that what the CLI does?
29/08/2006 21:39:53 ‹Edward2› I'll take a look
29/08/2006 21:39:55 ‹James› no, i thin SB added the queue into his app
29/08/2006 21:40:14 ‹James› the CLI is just a display now type setup
29/08/2006 21:40:19 ‹Edward2› Wish I had the source to that
29/08/2006 21:40:21 ‹KevinH› SB's has a three factor queue system... message priority, timed display and each player individualy
29/08/2006 21:40:30 ‹Edward2› pfffff
29/08/2006 21:40:36 ‹Edward2› more work
29/08/2006 21:40:43 ‹KevinH› yep - it was quite a lot of work I remember
29/08/2006 21:40:51 ‹KevinH› but very useful
29/08/2006 21:41:20 ‹KevinH› another plea for source code then for SB !
29/08/2006 21:41:34 ‹KevinH› would save a ton of work
29/08/2006 21:41:44 ‹James› very useful.. as thinks like xAP News send almost a stream of news flashes and allowing a high priority callerid message to display now is a real requirement
29/08/2006 21:42:02 ‹Edward2› I never really played with SB's connector - I do have a binary so I can try to replicate the behaviour of that
29/08/2006 21:42:24 ‹Edward2› OK, OK
29/08/2006 21:42:28 ‹Edward2› :-)
29/08/2006 21:42:48 ‹Edward2› Really enjoying the Perl - really good with xAP
29/08/2006 21:43:03 ‹KevinH› We did talk around at one stage pulling the 'queue' bit out into a separate application so that lots of xAP apps could benefit as it's a useful bit of code .
29/08/2006 21:43:31 ‹James› well i have that code in vb6 as i use it in the intranet launcher
29/08/2006 21:43:51 ‹Edward2› Well yes, I can see the usefulness of a centralised message display coordinator
29/08/2006 21:44:21 ‹James› although centralised is not really the xAP way ;-)
29/08/2006 21:44:31 ‹KevinH› J .. a priority based, multi destination queue with timed delays per message ?
29/08/2006 21:44:59 ‹James› well it's only single destination, but it does the rest
29/08/2006 21:45:02 ‹KevinH› But neither is lots of disparate features within one application
29/08/2006 21:45:39 ‹KevinH› To an extent this 'queue' could be a scheduler too
29/08/2006 21:45:52 ‹KevinH› or a timed delay type thing
29/08/2006 21:46:39 ‹James› so who's goin to make it then ;-)
29/08/2006 21:46:48 * KevinH quit
29/08/2006 21:46:52 * KevinH joins My room
29/08/2006 21:48:18 * kevint joins My room
29/08/2006 21:48:45 ‹KevinH› hope we didn't miss anything KT ;-)
29/08/2006 21:49:03 ‹KevinH› J - diod you fix anything on the biz domain ?
29/08/2006 21:49:41 ‹James› no, reported the fault, but didn't hear anything
29/08/2006 21:50:18 ‹KevinH› did other people have issues then - I saw the comment about emails
29/08/2006 21:50:42 ‹James› it blipped for be for about 4 hours. thats all
29/08/2006 21:50:43 ‹KevinH› The hosts workaround dunt work for me
29/08/2006 21:50:54 ‹James› o
29/08/2006 21:51:43 ‹kevint› sorry back again
29/08/2006 21:52:13 ‹KevinH› I thought I had a great workaround for HV's named I/O as it's all stored (and readable) from HV via teh serial port so i didnt have to have it local
29/08/2006 21:52:32 ‹KevinH› works great for events and infos
29/08/2006 21:52:51 ‹KevinH› but forgot wildcard matching :-(
29/08/2006 21:53:12 ‹kevint› so did we have a consensus on the schema for analog values, New schema, BSC extension or new schema based on BSC..?
29/08/2006 21:53:28 ‹kevint› or more discussion required
29/08/2006 21:53:52 ‹KevinH› I'm sort of thinking that we couldn't sensib;y squeeze it in within BSC without compromising it - or both
29/08/2006 21:54:15 ‹James› my vote is "new schema based on BSC"
29/08/2006 21:54:41 ‹KevinH› so maybe need a new schema - not based on BSc persay but on the same methodology 'status'/'event'/'cmd'
29/08/2006 21:55:27 ‹James› that what i imply from based on bsc
29/08/2006 21:55:32 ‹kevint› I would go with that, BSC is really too core to muck about with
29/08/2006 21:55:48 ‹KevinH› In a way deriving thinsg from BSc is useful though - means perhpas taht existing precious code in PIC devices (thinking netiom) can be used and adapted
29/08/2006 21:56:06 ‹KevinH› I know space is tight there
29/08/2006 21:56:15 ‹kevint› so thats 3 votes out of six
29/08/2006 21:56:28 ‹KevinH› and that he wants to add 1-wire/temp sensors to the next version
29/08/2006 21:56:29 ‹kevint› I mean 5 edward ;)
29/08/2006 21:56:30 ‹James› so DSC is good,, only 1 byte change
29/08/2006 21:56:56 ‹KevinH› TSC - telemetry ?
29/08/2006 21:57:00 ‹kevint› "instrumentation" is definitely too long !
29/08/2006 21:57:20 ‹James› it's official so itll have to start xAP
29/08/2006 21:58:08 ‹kevint› I like xAPTSC
29/08/2006 21:58:43 ‹James› sold
29/08/2006 21:58:53 ‹KevinH› I think Gregg and Lehane have great input so far and could thrash out a schema ?? - and you too KT if you are game for it ?
29/08/2006 21:59:22 ‹kevint› Happy to contribute
29/08/2006 21:59:37 ‹kevint› Now about that Watchdog app......
29/08/2006 21:59:41 ‹kevint› ;)
29/08/2006 22:00:10 ‹James› ok, i'll do that one
29/08/2006 22:00:17 * Edward2 quit (timeout)
29/08/2006 22:00:22 ‹KevinH› The problem is that these things always seem so simple at the start and little things crop up taht make it ever more complex
29/08/2006 22:00:49 ‹James› BSC took months
29/08/2006 22:00:55 ‹kevint› FP - Fink Positive
29/08/2006 22:01:00 ‹James› but it's good when its done
29/08/2006 22:01:01 ‹KevinH› J we talked a while back too about advertising configuration pages (web absed) via Internet launcher
29/08/2006 22:01:09 ‹kevint› This wont be quick but we can make a start
29/08/2006 22:01:25 ‹James› yep, didn't get very far with it
29/08/2006 22:01:57 ‹KevinH› just one for the list as it tidies thinsg up a bit
29/08/2006 22:02:12 * Edward3 joins My room
29/08/2006 22:02:14 ‹kevint› That sounds similar to something we talked about last year
29/08/2006 22:02:30 ‹KevinH› If I can squeeze teh web config in teh C-Bus and HV controllers I'll make the intranet compatible
29/08/2006 22:02:36 * Edward quit (timeout)
29/08/2006 22:02:58 * Edward joins My room
29/08/2006 22:03:02 ‹James› the slight snag with the linked configs is dealing with apps that oonly allow local config
29/08/2006 22:03:06 ‹kevint› where I suggested that some intranet apps were not in the "User Tier"
29/08/2006 22:03:24 ‹KevinH› E - we must talk re lighting schema too
29/08/2006 22:03:32 ‹James› yep, and grouping etc etc
29/08/2006 22:03:44 * Edward3 quit (timeout)
29/08/2006 22:04:05 ‹KevinH› you mean apps that dont have a web config at all ?
29/08/2006 22:04:28 ‹Edward› I know, I know - I need to give up my day job
29/08/2006 22:04:54 ‹James› no some apps like FP and SB have an option so that the main web pages are available from anywhere but the config pages only work from localhost
29/08/2006 22:05:09 ‹kevint› I mean intranet apps which are not realy for the User but have a web interface soley for management & config
29/08/2006 22:05:10 ‹KevinH› That's the big chestnut we all know is lurking still.. xAP based config - ahh I see
29/08/2006 22:05:28 ‹KevinH› Is there a reason they enforce that ? aside from security /
29/08/2006 22:05:37 ‹James› just security
29/08/2006 22:05:51 ‹KevinH› back in 2
29/08/2006 22:07:22 ‹kevint› Edward, is 2.6.3 the current working version of the x10 connector?
29/08/2006 22:08:37 * kevint joins My room
29/08/2006 22:09:31 * kevint joins My room
29/08/2006 22:11:13 ‹KevinH› I have 2.6.3.2 KT but that might be one E is testing with wildcard changes
29/08/2006 22:12:27 ‹KevinH› KT - did you see that post re embedded space on xAP Auto .....
29/08/2006 22:13:11 ‹kevint› ah OK
29/08/2006 22:13:24 ‹kevint› embedded space, I will look
29/08/2006 22:14:44 ‹Edward› (sadly) I can't remember - goes off to look...
29/08/2006 22:15:44 ‹KevinH› KT it's in 'developer' white space
29/08/2006 22:17:56 ‹Edward› 2.6.3.2 (debug) is the latest development version with wildcard fixes - 2.6.3 the latest release
29/08/2006 22:18:43 ‹kevint› Thanks
29/08/2006 22:19:14 ‹Edward› Had a chance to test 2.6.3.2 KH?
29/08/2006 22:20:01 ‹kevint› Looking at the white space thing, the spaces are in the messages supplied with the messend app?
29/08/2006 22:20:25 ‹kevint› These messages were basically taken from Viewer I think
29/08/2006 22:21:05 ‹kevint› I will look at the packaged messages more closely
29/08/2006 22:21:46 ‹KevinH› I have played briefly E but I haven't gone through all the permutations I ddd before - your explanation of why it was not wildcarding as expected seemed to make sense
29/08/2006 22:21:54 ‹kevint› but I didnt think there was a prblem with 'text<space>=<space>1'
29/08/2006 22:23:01 ‹KevinH› The value of text would be <space>1 and not 1 though - ie 2 characters
29/08/2006 22:24:10 ‹kevint› this isnt back to putting numbers in text fields is it ;)
29/08/2006 22:25:00 ‹kevint› perhaps I should have said value<space>=<space>1
29/08/2006 22:26:40 ‹Edward› OK - it seemed good to me so I'll go for an X10 2.6.4 release
29/08/2006 22:27:23 ‹KevinH› You aren't going to Net2 it are you - that's my current issue (disconnects) ?
29/08/2006 22:27:40 * lehane joins My room
29/08/2006 22:28:01 ‹Edward› That'll be 2.7
29/08/2006 22:28:43 ‹KevinH› it isnt related to the parameter being text - it's just that all leading and trailing space in xAP parameter values IS significant
29/08/2006 22:29:27 ‹KevinH› Hi Lehane
29/08/2006 22:29:35 ‹lehane› Hi all
29/08/2006 22:30:14 ‹James› hi
29/08/2006 22:30:15 ‹KevinH› We think you and Gregg are making greta headway on a sensor type schema
29/08/2006 22:30:45 ‹lehane› I'm just trying to get the ball rolling.....
29/08/2006 22:30:51 ‹KevinH› and soncensus seems to be not to try for backward compatability with BSC
29/08/2006 22:31:00 ‹KevinH› concensus
29/08/2006 22:31:30 ‹lehane› Yes, probably should standalone
29/08/2006 22:32:02 ‹kevint› KH please pop up on skype
29/08/2006 22:32:05 ‹KevinH› but to create a new schema similar in style and mechanism to BSC - perhpas called xAPTSC for 'telemetry' with events,infos and commands
29/08/2006 22:32:47 ‹KevinH› hoping that the people using PIC's can leverage their existing realtime parsing to support both scema
29/08/2006 22:33:14 ‹lehane› OK. I guess most PIC's will be source
29/08/2006 22:34:08 ‹KevinH› I know Roy at Phedrus wants to add 1-wire support to his next xAP netiom so he would use the xAPTSC for say temp sensors
29/08/2006 22:36:36 ‹Gregg› Are there opinions on having multiple blocks to correspond to multiple sensed values from a single device? I think Lehane and I are of the opinion that should be allowed (Lehane please correct if not)
29/08/2006 22:38:32 ‹lehane› yes - like temp and humidity
29/08/2006 22:39:42 ‹KevinH› Multiple blocks are permitted in xAP - we sort of tried to bundle context within one block wherever possible though - the issue comes where you are using subaddresses - I am assuming you would then have 1 sub address for a compound sensor would you ?
29/08/2006 22:40:00 ‹Gregg› yes
29/08/2006 22:40:30 ‹Gregg› and, for 1-wire, it would map directly to the 1-wire serial ID
29/08/2006 22:40:47 ‹kevint› sorry do u mean that temp and humidty values in the same device would have the same address?
29/08/2006 22:41:10 ‹Gregg› subaddress--yes
29/08/2006 22:41:29 ‹KevinH› What I was meaning is that if you report two blocks within one message they woudl effectively be both from one sub address
29/08/2006 22:41:37 ‹Gregg› yes
29/08/2006 22:42:46 ‹KevinH› Let em think on that a sec - seems reasonable if several 1 wire devices might be served by one app - each with their own sub address
29/08/2006 22:43:26 ‹Gregg› I happen to be working this moment on one--hence the interest
29/08/2006 22:44:33 ‹KevinH› If there was an output 1 wire device - could it have two embedded sensors that were each settable separately ?
29/08/2006 22:44:56 ‹KevinH› not sensors - say a D to A or something
29/08/2006 22:45:33 ‹Gregg› I believe that there are 1-wire counter devices that support 2 counters concurrently. Something like that?
29/08/2006 22:45:44 ‹KevinH› Output devices ?
29/08/2006 22:46:59 ‹lehane› I haven't given as much thought to outputs
29/08/2006 22:47:05 ‹Gregg› I'm unaware of any. But, does it matter?
29/08/2006 22:47:11 ‹KevinH› The way most people 'fudge' messages with 2 blocks is to process them as two separate messages immediately after each other, not absolutely technically correct but it's a close solution
29/08/2006 22:48:12 ‹Gregg› yes--but a pain as you're then talking about have to map a single real device into 2 subaddresses/UIDs--right?
29/08/2006 22:48:19 ‹KevinH› when you send something it isnt possible to do this as one message might be missed so obviously you have to use 2 bodies - in say a .cmd message
29/08/2006 22:49:30 ‹Gregg› sorry--thought you meant that people should be splitting the output as 2 messages rather than processing them as 2
29/08/2006 22:50:02 ‹KevinH› I agree - from input devices its not such an issue but if your commanding them to change state then it's more crucial. Thinking on this though... if they were 2 separate UID's and you needed to chnage 2 things simultaneously then they would have to be wildcard addressable as one - which si awkward
29/08/2006 22:50:31 ‹KevinH› I think combining them as one is logical
29/08/2006 22:50:48 ‹KevinH› . or at least a design choice thing
29/08/2006 22:51:38 ‹KevinH› What I would like is for this new schema to be as simple as possible to implement but not to compromise its capabilities
29/08/2006 22:51:44 ‹lehane› In most cases processing blocks sequentially is of little consequence - unless you are into demolitons
29/08/2006 22:52:07 ‹KevinH› Yes - I agree
29/08/2006 22:53:09 ‹KevinH› Sending two separate mesages to action a result when both MUSTchange state is not ideal though
29/08/2006 22:54:14 ‹KevinH› incoming parsing of sequential blocks is very acceptable because you know you do have all the related content
29/08/2006 22:56:00 ‹KevinH› If you have two output devices on one subUID that can be individually controlled then the particular device you are addressing in say a .cmd has to be identified in the message body
29/08/2006 22:56:31 ‹KevinH› either by its section name or by an identifying parameter
29/08/2006 22:56:31 ‹Gregg› agreed--and the implication is probably that the cmd must target the subaddress--right?
29/08/2006 22:57:20 ‹KevinH› one thing that is very difficult is when you have a body parameter that has a context set by other parameters
29/08/2006 22:57:50 ‹KevinH› device=waterlevel level=100
29/08/2006 22:58:01 ‹KevinH› device=temp level=100
29/08/2006 22:58:37 ‹KevinH› now - a controller needs to know that the 'device' parameter tells it what the level reading is for
29/08/2006 22:59:14 ‹KevinH› and nothing makes that apparent in the message - only the schema definition conveys this
29/08/2006 22:59:39 ‹KevinH› so controllers need access to schema info electronically to be able to interpret values
29/08/2006 23:00:21 ‹KevinH› This happens a lot in xPL but not at all in BSC say - although it is present in some schema
29/08/2006 23:00:40 ‹Gregg› not much different in concept from the use of "ID" in BSC cmd messages is it?
29/08/2006 23:00:42 ‹KevinH› (some xAP schema)
29/08/2006 23:01:09 ‹KevinH› But ID is effectively repeated in the target line (always)
29/08/2006 23:01:41 ‹lehane› James - testing getfromraw - can't seem to extract the name/value but rawmessage string contains data OK
29/08/2006 23:02:01 ‹Gregg› ?? I must be missing something; I'm referring to a BSC cmd that is targeting the device--not a subaddress
29/08/2006 23:02:40 ‹James› L, can you give me the line from your script?
29/08/2006 23:02:47 ‹KevinH› Ahh - targeting the whole application ..
29/08/2006 23:03:03 ‹Gregg› yes--we were talking about a multi-block example--right?
29/08/2006 23:03:27 ‹Gregg› I'm suggesting that that model is much different than what you were describing above
29/08/2006 23:03:43 ‹Gregg› is not much different
29/08/2006 23:04:07 ‹lehane› xap.getfromraw(rawmessage, "event, state";) and the rawmessage = event,=~~event,time=22:58:55~~event,state=0~~event,type=faulted~~event,zone=6~~
29/08/2006 23:04:13 ‹KevinH› I thought you were advocating multiblock from one UID - eg a 1 wire that had a humidity and temp sensor but was reporting as 1 sub address
29/08/2006 23:04:23 ‹Gregg› I am
29/08/2006 23:04:36 ‹KevinH› not a main address ..
29/08/2006 23:04:42 ‹James› with or without that space after teh comma
29/08/2006 23:04:44 ‹James› ?
29/08/2006 23:04:54 ‹KevinH› ie a UID that deosnt end in 00
29/08/2006 23:04:57 ‹Gregg› I'm suggesting that the multiblock BSC cmd targeting a main address is much like a multiblock "TSC" cmd targetting a subaddress
29/08/2006 23:05:52 ‹Gregg› hence, my comment that "device" is much like "ID" given that kind of a model
29/08/2006 23:05:55 ‹KevinH› A BSC.cmd never targets a main address... just checking this...
29/08/2006 23:06:07 ‹KevinH› it targets 'all sub addresses'
29/08/2006 23:06:12 ‹KevinH› doesnt it
29/08/2006 23:06:21 ‹lehane› J - yep ££$£$"!¬ spaces! Fixed
29/08/2006 23:06:30 ‹James› yay
29/08/2006 23:06:45 ‹Gregg› yes--but not much different IMO
29/08/2006 23:06:53 ‹lehane› Now I have the caddx PIR's controlling lighting
29/08/2006 23:07:37 ‹James› you'll need to write it all up for the website now ;-)
29/08/2006 23:08:05 ‹lehane› need a login <hint>
29/08/2006 23:08:26 ‹James› on it
29/08/2006 23:09:44 ‹KevinH› The ID is unique and related intrinsically to a sub address ID=56 is subaddress xxxxxx56 if you name something 'temperature' now its vague as to what it relates to
29/08/2006 23:10:53 ‹KevinH› and how you would target a mesage back to control it. I suppose what I'm saying is that BSC defined ID as being a numbered link parameter with the UID whereas ..
29/08/2006 23:11:26 ‹James› L, you have mail
29/08/2006 23:11:54 ‹KevinH› device=humidity is more general . Maybe BSCID SubUIDwould have bene even better in hindsight
29/08/2006 23:12:04 ‹Gregg› agree--that it is more formal; what I disagree w/ is that the controller now has to know more wrt something like "device" than the concept of "id"
29/08/2006 23:13:07 ‹KevinH› Ahh we're at crosspurposes a little here I think...
29/08/2006 23:13:49 ‹lehane› Consider a TOM10 - under TSC that is 10 devices
29/08/2006 23:14:32 ‹KevinH› If you define 'device' in the schema as having this interpretation then that;'s fine
29/08/2006 23:15:49 ‹KevinH› what I was saying as a generalisation - not TSC specific is that when you have some parameters in one block that set a context as to the meaning of others it makes it impossible for a controller to track state of that device without specific schema knowledge
29/08/2006 23:16:26 ‹KevinH› deviceCode=4 Housecode=D State=On for example
29/08/2006 23:16:47 ‹KevinH› Housecode=D State=All units on
29/08/2006 23:17:13 ‹KevinH› make it very difficult to tracks the state of D4
29/08/2006 23:17:18 ‹KevinH› and D5
29/08/2006 23:18:34 ‹KevinH› .. Which is awkward for things like MisterHouse and HomeSeer and xAp Floorplan unless they are preprogrammed to undertsand the schema or can get at an electronic definition of teh schema
29/08/2006 23:19:13 ‹Gregg› fwiw: all of misterhouse's implementation *does* understand the schem
29/08/2006 23:19:36 ‹KevinH› Yes - because you code it all in :-)
29/08/2006 23:20:12 ‹Gregg› in the context of mh--it wouldn't make sense to do otherwise
29/08/2006 23:21:09 ‹lehane› But which implementation could be expected to pick up a new schema and interpet the content without coding?
29/08/2006 23:21:50 ‹KevinH› Should someone introduce their own schema for say 'washing machine control' it means that MH can't interact with that device easily - nor necessarily can it even display raw values without some ability for teh user to tell it which are the parameters that establish context
29/08/2006 23:22:22 ‹KevinH› If we had an electronic repository of schema it would be far more achievable
29/08/2006 23:22:30 ‹Gregg› yes--that's entirely true--by design
29/08/2006 23:23:00 ‹lehane› only with some pretty rigourous control over name/vaule pairs
29/08/2006 23:23:06 ‹KevinH› and if we minimise 'context' being established by multiple parameters within a block
29/08/2006 23:23:50 ‹Gregg› agree w/ the concept of minimize; just not to such an extreme that no (appropriate) exceptions are permitted
29/08/2006 23:24:07 ‹KevinH› Its not at all easy to do this ...
29/08/2006 23:24:24 ‹KevinH› many schema already have this - like wether etc
29/08/2006 23:24:54 ‹KevinH› if you dont re-use parameter names it helps too
29/08/2006 23:25:35 ‹KevinH› for example sensor=humidity value=80 is not so helpful as sensor=humidy hvalue=80
29/08/2006 23:26:24 ‹Gregg› I think Lehane had used "type" to address the context for value
29/08/2006 23:26:26 ‹KevinH› as hvalue is now unique and can be plucked and displayed/utilised in a controller application
29/08/2006 23:26:56 ‹KevinH› it is definitively the humidty value
29/08/2006 23:26:57 ‹lehane› only if the units are standardised - which is where this started with temperature...
29/08/2006 23:27:36 ‹KevinH› so could we have mutiple blocks with different parameter names or at least different block headers
29/08/2006 23:27:56 ‹Gregg› I had assumed different block headers
29/08/2006 23:28:03 ‹KevinH› humidty.report
29/08/2006 23:28:10 ‹lehane› yep
29/08/2006 23:28:34 ‹KevinH› ahh well thats great because we've taken the context outside the block
29/08/2006 23:29:10 ‹lehane› so we can drop the type = (or sensor =)
29/08/2006 23:29:40 ‹KevinH› yes - you may find you want min max and averages reported though
29/08/2006 23:30:04 ‹lehane› I'd still like the units (Celsius, Amps, whatever)
29/08/2006 23:30:06 ‹KevinH› and its better (again) not to say report=max value=100
29/08/2006 23:30:25 ‹Gregg› cannot those be reported as distinct elements w/i the block?
29/08/2006 23:30:31 ‹KevinH› but to have valuemax valuemin etc
29/08/2006 23:30:34 ‹KevinH› yep
29/08/2006 23:30:51 ‹Gregg› L--agree w/ the units
29/08/2006 23:31:09 ‹lehane› prob with max min avg is what time period and reset?
29/08/2006 23:31:10 ‹KevinH› The units is something that I cant see logically going anyhwere else except withina block
29/08/2006 23:31:35 ‹KevinH› these things start easy and get involved don't they
29/08/2006 23:31:59 ‹lehane› oh yes...
29/08/2006 23:32:26 ‹KevinH› you might for example issue a 'cmd' to the app asking for an average over a time you define in that command
29/08/2006 23:33:03 ‹lehane› my view is the consuming app should do the avg over what time period is useful to the proces/device it controls
29/08/2006 23:33:06 ‹KevinH› we do have a standard xAP date/time field so you could report that as start/end in teh avg/max/min messages
29/08/2006 23:34:28 ‹KevinH› or have a dailymax dailymin hourlymax hourlymin ... trouble is everyone will want something different
29/08/2006 23:34:37 ‹lehane› Example for a roomstat-avg for radiator controller may be 10 mins but avg for stats reporting may be 24hrs
29/08/2006 23:35:16 * kevint quit (timeout)
29/08/2006 23:35:28 ‹Gregg› L-agree that the consuming app should be allowed; but should also not prevent the sensor from reporting it as the real values may be necessary to get an accurate average
29/08/2006 23:35:57 ‹lehane› and asking a PIC to respond to a specific time period request may be too much for it
29/08/2006 23:36:03 ‹lehane› and asking a PIC to respond to a specific time period request may be too much for it
29/08/2006 23:36:26 ‹KevinH› indeed - I suppose there's no reason not to take the same approach as BSC - allow such parameters to be added into the block - enabling the author to report extra functionality if there device can support it
29/08/2006 23:36:39 ‹lehane› yes
29/08/2006 23:36:47 ‹Gregg› agreed
29/08/2006 23:36:52 ‹KevinH› optional parameters - just reserve some structure and parameter names for obvious uses
29/08/2006 23:37:32 ‹KevinH› dailymax or whatever and allow a start/end time to be reported for weird ranges
29/08/2006 23:37:52 ‹Gregg› sorry to leave; but being called from above ...
29/08/2006 23:38:07 ‹KevinH› Nite
29/08/2006 23:38:14 ‹lehane› OK, its getting late - shall I put together another rev and some examples, taking into account all the above?
29/08/2006 23:38:21 * Gregg quit
29/08/2006 23:38:22 ‹James› nite
29/08/2006 23:38:33 ‹KevinH› yep - sounds great - I think KT has some input too
29/08/2006 23:39:05 ‹KevinH› One quick Q - is your DMX lighting schema fairly sorted now ?
29/08/2006 23:39:53 ‹KevinH› Would be good to get C-Bus, DMX and Dynalite etc all on one schema
29/08/2006 23:40:07 ‹KevinH› although I must admit I use BSC mostly here
29/08/2006 23:40:16 ‹lehane› I'm happy to change it - mostly it is BSC
29/08/2006 23:40:24 ‹KevinH› me too ;-)
29/08/2006 23:40:40 ‹lehane› but a lighting schema would be good...
29/08/2006 23:40:49 ‹lehane› areas, scenes, etc
29/08/2006 23:40:50 ‹KevinH› we could agree an additional ramp parameter usage perhpas in teh BSc bodies
29/08/2006 23:41:05 ‹KevinH› ramprate
29/08/2006 23:41:39 ‹KevinH› Areas, Groups , Scenes is one other thing that starts simple and has big overlaps in not jsut lighting but all xAP addressing
29/08/2006 23:41:40 ‹lehane› units? % per sec, time from x to y?
29/08/2006 23:42:24 ‹KevinH› Mine is just the rate at which it will change from X to Y (number of secs)
29/08/2006 23:42:47 ‹lehane› x=current, y=target
29/08/2006 23:43:32 ‹KevinH› having said that maybe its the rate it will change from 0 to 100% - ie the ramp is constant and a change from 10 to 20% is achieved in half the time 10 to 30% is - must check that
29/08/2006 23:44:22 ‹lehane› OK, I must go too...
29/08/2006 23:44:37 ‹KevinH› ..and I only ramp from 'current' to new so I specify a destination level= and never need a start value - maybe I should have one
29/08/2006 23:45:01 ‹KevinH› Nite there...
29/08/2006 23:45:07 ‹lehane› nite all
29/08/2006 23:45:10 * lehane quit
29/08/2006 23:45:55 ‹James› to change the subject, i may be close to my first commercial app :-)
29/08/2006 23:46:17 ‹KevinH› ?
29/08/2006 23:46:24 ‹James› might need a few opinions in a bit
29/08/2006 23:46:38 ‹James› well all my apps are free for non-commercial use
29/08/2006 23:46:57 ‹KevinH› was looking for a money symbol
29/08/2006 23:47:41 ‹KevinH› It would be good to get some rewards
29/08/2006 23:48:01 * James quit (timeout)
29/08/2006 23:48:34 ‹KevinH› People think I make money on teh xAP gateways but it's cost me far more than I've made - or ever will unless I hike the price
29/08/2006 23:49:14 * James joins My room
29/08/2006 23:49:29 ‹James› arg, adsl modem died
29/08/2006 23:49:48 ‹KevinH› does it do that often ?
29/08/2006 23:49:55 ‹James› ping 3 is the app as it's getting quite powerful
29/08/2006 23:50:19 ‹James› force9 stuffed me onto a llu deal with tiscalli and its glitchy
29/08/2006 23:50:21 ‹KevinH› and useful to many large companies/organisations
29/08/2006 23:51:08 ‹James› yep. there will be licenses for different numbers of scans, and also there will be a displayscreen app for an on the wall status display
29/08/2006 23:51:31 ‹KevinH› hope it does well for you
29/08/2006 23:51:39 ‹James› free licenses for testers ;-)
29/08/2006 23:52:16 ‹James› who knows, but as i don't currently allow commercial use of it at all then i do need a route
29/08/2006 23:52:39 ‹KevinH› I have the C-Bus HV thing mostly working now although I think there may be a little issue in teh latest HV roms - but it could be just me
29/08/2006 23:53:07 ‹KevinH› I get the wrong values on teh event updates
29/08/2006 23:53:49 ‹James› we should team up and make a xAP shop
29/08/2006 23:53:50 ‹KevinH› so when I get an event I have to read all the current values again (256 lights) which is a pain over a serial link
29/08/2006 23:54:07 ‹KevinH› In Barnsley
29/08/2006 23:54:09 ‹James› must take ages
29/08/2006 23:54:18 ‹James› a virtual shop
29/08/2006 23:54:44 ‹KevinH› Its just awkward to receive so much data reliably without losing other things
29/08/2006 23:55:16 ‹James› as you don't have a huge buffer i guess
29/08/2006 23:55:35 ‹KevinH› and HV is really one sided - it will send anythinga ta naythime in any quantity - but will only recieve tiny bits and you ahve to wait for acknowledges before teh enxt few characters
29/08/2006 23:56:01 ‹KevinH› but at least it can type better than me
29/08/2006 23:56:57 ‹KevinH› Will your work licence Ping3 ?
29/08/2006 23:57:13 ‹James› i would have to ask the boss
29/08/2006 23:57:22 ‹James› :-)
29/08/2006 23:57:26 ‹KevinH› did you use work time to develop it ?
29/08/2006 23:57:37 ‹James› noo
29/08/2006 23:57:49 ‹KevinH› well you're on solid groud then
29/08/2006 23:58:01 ‹James› and they got free use of it too for it's trial
29/08/2006 23:58:33 ‹James› most of pings features were from work needs, like the html writing and the command to let it upload toa website
29/08/2006 23:59:27 ‹KevinH› Sure it'll get on their requisition list then
30/08/2006 00:00:35 ‹KevinH› Arrrgh I hate X10....
30/08/2006 00:00:56 ‹James› well if i only sell the odd one then its still worth it
30/08/2006 00:02:15 ‹KevinH› The level encoding in byte values is awful - even in teh docs it says - 'rather than try to explain the format here's some VB code that was written to return the level from a byte -a nd 30 odd lines of code
30/08/2006 00:03:57 ‹KevinH› every one you sell is profit - unlike hardware that has a build cost..
30/08/2006 00:04:00 ‹James› so why are you doing x10.. does HV do that for you or do you talk raw x10 to hv
30/08/2006 00:04:32 ‹KevinH› It am trying to xAP enable all HV's I/O via BSC
30/08/2006 00:04:46 ‹KevinH› and also using teh X10 schema
30/08/2006 00:05:07 ‹KevinH› it talks 'events' to me but reports 16 devices at a time
30/08/2006 00:05:21 ‹KevinH› and they are byte encoded
30/08/2006 00:05:48 ‹James› andd it use x10 protocol over the serial from the hv?
30/08/2006 00:06:39 ‹KevinH› levels, state and more all embedded in one byte and then stupidly having saved all that space its preceeded by "X10HCUPDATE:" ;-)
30/08/2006 00:06:55 ‹KevinH› no HV messages
30/08/2006 00:06:59 ‹James› lol
30/08/2006 00:07:03 ‹James› k
30/08/2006 00:07:49 ‹KevinH› Even worse are teh varaiable updates - HV has 256 variables and all their values are reported at once - even if only one chnages
30/08/2006 00:08:21 ‹KevinH› - and if you include a variable in a loop say as a counter you get all 256 reported every loop
30/08/2006 00:08:32 ‹KevinH› include=update
30/08/2006 00:08:52 ‹James› so the serial link it busy all the time then
30/08/2006 00:09:02 ‹James› how fast is it? 19200
30/08/2006 00:09:10 ‹KevinH› yes
30/08/2006 00:09:23 ‹KevinH› can be faster on HV pro I think
30/08/2006 00:09:35 ‹James› well i suppose dmx is only 9600 and it does 512 channels
30/08/2006 00:09:57 ‹KevinH› You have to learn not to put variables in a loop and to only update variables if you want them to change
30/08/2006 00:10:08 ‹KevinH› Put the update in an IF statement
30/08/2006 00:10:26 ‹James› and end up running out of code space
30/08/2006 00:10:26 ‹KevinH› dont keep saying VAR2=0 - as that's an update too
30/08/2006 00:10:51 ‹KevinH› very inefficient
30/08/2006 00:11:53 ‹James› so how many hurdles are left until its done?
30/08/2006 00:11:58 ‹KevinH› I hope no one uses P16 housecode
30/08/2006 00:12:09 ‹KevinH› as it's gone
30/08/2006 00:12:16 ‹KevinH› only 255 UID's
30/08/2006 00:12:50 ‹James› but then missing one in the middle may be safer
30/08/2006 00:12:55 ‹KevinH› The custom lighting is finished except for the named sub addresses
30/08/2006 00:13:03 ‹James› as i have my x10 lights on a and my switches on P
30/08/2006 00:13:29 ‹KevinH› I'm on B D and G
30/08/2006 00:13:30 ‹James› and i do indeed use p16 ;-)
30/08/2006 00:13:39 ‹KevinH› tough
30/08/2006 00:13:44 ‹James› but no HV so you're fine
30/08/2006 00:14:15 ‹James› i thkni went for p1 and mis clicked to p16 and just left it like it
30/08/2006 00:15:10 ‹KevinH› Interestingly I have found that housecode M is being used somehwere - maybe a thermostat I once set up that it continually scans
30/08/2006 00:15:44 ‹KevinH› Also iHV has macros it runs when it sees X10 change state
30/08/2006 00:16:14 ‹KevinH› yet it you ask it to change the stae of an X10 device it then doesnt run teh macro
30/08/2006 00:16:37 ‹KevinH› only if something external makes the change and it sees it on teh line
30/08/2006 00:17:12 ‹KevinH› I need to see if I can squeeze in a web config/browser still
30/08/2006 00:17:30 ‹KevinH› prepared to take xAP out for that I guess
30/08/2006 00:17:32 ‹James› i did ahve a while when i told a1 to turn on and both a1 and a2 turned on
30/08/2006 00:17:50 ‹Edward› nite nite
30/08/2006 00:18:09 ‹KevinH› I've seen similar to that -a nd I had one where D4 ON worked but D4 off turned D6 ON
30/08/2006 00:18:13 ‹KevinH› nite E
30/08/2006 00:18:54 ‹James› nite
30/08/2006 00:19:07 ‹KevinH› I must get some sleep too
30/08/2006 00:19:34 ‹James› me too, spend too much time tweaking ping3 at the mo
30/08/2006 00:20:03 ‹KevinH› well , good luck and I expect the release by teh weekend
30/08/2006 00:20:25 ‹KevinH› nite
30/08/2006 00:20:29 ‹James› nite