These are chat archives for nightscout/intend-to-bolus

9th
Sep 2015
Scott Leibrand
@scottleibrand
Sep 09 2015 08:58
@audiefile would love to get you to the point where you're using the same loop.sh core code as @amazaheri @danamlewis and @esteward so all our efforts can be pushing the same thing forward and we all benefit from all the improvements (especially as we start to replace components with @loudnate and @channemann's stuff)
Ben West
@bewest
Sep 09 2015 18:50
I'm less worried about the hand authored loops atm, we could still use another dozen in our "garden" to learn from
eventually we'll have enough info to create a tool that manages the loop automatically
Ali Mazaheri
@amazaheri
Sep 09 2015 18:56
:+1:
@bewest so I thought RileyLink was programmed so we still need to program it when you are in town?
Ben West
@bewest
Sep 09 2015 18:57
when I was at 'nate's we flashed all four
one to mark wilson, one for you, two for me :-)
Ali Mazaheri
@amazaheri
Sep 09 2015 19:01
so u still need the tools ?
or can we just connect via BT :)
Ben West
@bewest
Sep 09 2015 19:02
should be able to connect right away
you'll need tool to get updates
Ali Mazaheri
@amazaheri
Sep 09 2015 19:02
cool
order it today super excited
Rachel Sandlain
@audiefile
Sep 09 2015 19:46
@scottleibrand I'll take another look at it. I remember it didn't run but not why. I think there were calls to commands I didn't have and I know I had issues with setting up aliases so that they would actually run
Ben West
@bewest
Sep 09 2015 19:54
yeah
while decocare is now smarter wrt to model differences, openaps is still clueless... doesn' know that you can't read_bg_targets on 512 or whatever
for those, need a use that can generate from config/args to simulate similar thing
Ali Mazaheri
@amazaheri
Sep 09 2015 19:57
I will send a PR with sample ones I created using 722 and share with folks, may be we can include in the docs
for now
if u use x12 download these, change the values and use in the loop
Ben West
@bewest
Sep 09 2015 19:59
or, we develop special vendor/use for those use pump legacy_bg_targets --high 180 --low 130 and similar for profile
Ali Mazaheri
@amazaheri
Sep 09 2015 19:59
even better
Chris Hannemann
@channemann
Sep 09 2015 20:13
I did a little science experiment to check the range of the CareLink that may be of interest; just posted the results here.
I'm guessing someone else has done similar tests before, so I'd love to hear what others have found
Ben West
@bewest
Sep 09 2015 20:24
fascinating
no one's done such a good job
nice
I would have expected 8 to be best
but you found 7 was best
orientation clearly matters
AND, the orientation that works best is not intuitive
is 13 like 8 or 7?
wow that is admirable work :-)
Chris Hannemann
@channemann
Sep 09 2015 20:28
thanks
13 is like 07
I was also surprised about the antenna facing away being best
I don't know much about radio transmission, so perhaps there is something obvious there
Ben West
@bewest
Sep 09 2015 20:36
what was underneath the wood floor?
Chris Hannemann
@channemann
Sep 09 2015 20:36
A whole bunch of 913 MHz repeaters, just like all wood floors
Ben West
@bewest
Sep 09 2015 20:36
when on it's side like in 7 and 8... hmm not sure shape of wave
basement? concrete?
Chris Hannemann
@channemann
Sep 09 2015 20:37
It's a ground floor apartment, no basement
Ben West
@bewest
Sep 09 2015 20:37
there's a 3d wave that propogates
Chris Hannemann
@channemann
Sep 09 2015 20:37
Probably concrete
Ben West
@bewest
Sep 09 2015 20:38
in mid-air, it at that orientation, the concrete foundation may interfere less with transmission
dunno
the backwards thing... no reason they couldn't mount the antenna upside down, I guess
the "back"/"front" is arbitrary
Chris Hannemann
@channemann
Sep 09 2015 20:39
Yeah, just how it's mounted
Ben West
@bewest
Sep 09 2015 20:40
hardware person laid pins out backwards or something
Chris Hannemann
@channemann
Sep 09 2015 20:40
Can't reverse it, though, 'cause there is a plastic standoff and the USB adapter is not symmetrically mounted in the z-plane
Ben West
@bewest
Sep 09 2015 20:40
happens all the time
right
makes it look prettier :)
Chris Hannemann
@channemann
Sep 09 2015 20:41
the most important thing!
Not that they would care about range for this
Ben West
@bewest
Sep 09 2015 20:42
I think it was originally conceived to be used in dr's office only
thinking was you'd go into dr, they'd download, analyze, and update your settings
Chris Hannemann
@channemann
Sep 09 2015 20:43
and often they drop it in an enclosure to minimize interference
Ben West
@bewest
Sep 09 2015 20:44
right, pre-512 they had lot of problems with shocking the pump, actually
so they built faraday cage into pump enclosure itself after 512
so that limits range
apparently the battery tube is connected to antenna
so you can increase surface area on the pump antenna by carefully wrapping foil, I guess
in the battery tube
Chris Hannemann
@channemann
Sep 09 2015 20:45
guh
Ben West
@bewest
Sep 09 2015 20:45
like putting pringles can in there ;-)
Chris Hannemann
@channemann
Sep 09 2015 20:46
Perhaps modify the outside of the battery itself?
Ben West
@bewest
Sep 09 2015 20:46
hehe, thought you'd like to know
no clue, I haven't tried it
impression I got is to wrap battery with foil, yeah
so outside long edges are foil wrapping making contact with battery enclosure
Chris Hannemann
@channemann
Sep 09 2015 20:49
There's the new business: selling pre-wrapped battery-antennas
Pete Schwamb
@ps2
Sep 09 2015 20:56
Yes, pump orientation relative to carelink antenna orientation will definitely make a difference. The emission often looks like a donut; reception along a line drawn through the hole of the donut will be particularly bad. The receiving antenna also has an axis along which it will be ‘deaf’ to signals.
Oskar Pearson
@oskarpearson
Sep 09 2015 20:59
Could I ask what the below actually means? What's "CalBGForGH", specifically?
{
    "body": "2d",
    "packet_size": 5,
    "amount": 45,
    "name": "CalBGForGH",
    "raw": "93-53-29-0f-2d",
    "date": "2015-09-09T19:19:00",
    "date_type": "minSpecific",
    "_tell": 45,
    "op": "0x0e"
  },
I'm seeing a very low glucose graphed with openaps-monitor, while in fact it was high
Ben West
@bewest
Sep 09 2015 21:00
GH = "glucose history"
meaning it comes from the glucose pages, not pump history pages
calbg is when you enter bg, looks like this was for a calibration to the medtronic cgm?
ah, donut shaped, @ps2 thanks. suspected the holes... so some space off-center is actually ideal?
also, the concrete foundation might be aiding the reflections, potentially?
Oskar Pearson
@oskarpearson
Sep 09 2015 21:10
Thanks. So the situation is that my wife had to leave the pump disconnected, and it lost the sensor. I only have the UK meter values mmol/l . When she got back, she had to re-connect the sensor and re-enter her current blood glucose. Her sugar was 16.7 (around 300). However, the graph / data seems to show her as being low (45).
It could be related to bewest/decoding-carelink#124 - but I'm really guessing
Chris Hannemann
@channemann
Sep 09 2015 21:17
Interesting, thanks @ps2
Pete Schwamb
@ps2
Sep 09 2015 21:19
Also, having the pump up against your body can often make it a lot more ‘quiet’ to the receiver.
Chris Hannemann
@channemann
Sep 09 2015 21:22
So buy a selfie-stick and attach it to my hip?
Pete Schwamb
@ps2
Sep 09 2015 21:22
hah :)
Ben West
@bewest
Sep 09 2015 21:23
yeah, one on each side
Chris Hannemann
@channemann
Sep 09 2015 21:23
But of course
@oskarpearson I think you are correct that it is the same (or similar) issue; not sure if/how @loudnate is getting around it at the moment, but I'm guessing it's by not calibrating when he's high (above 255?)
Dana Lewis
@danamlewis
Sep 09 2015 21:25
::is amused::
Ben West
@bewest
Sep 09 2015 21:26
we should have 3d-printed your dress, dana
Dana Lewis
@danamlewis
Sep 09 2015 21:26
that might be why (proximity to body) I have a hard time with it reading in the bed when it's clipped directly on my hip against my skin. Not sure if everyone else is letting it float, or if they're wearing it differently
Right?! :)
Oskar Pearson
@oskarpearson
Sep 09 2015 21:26
Ok, I'm starting to understand the issue a bit more. Thanks
Ben West
@bewest
Sep 09 2015 21:26
I would put it under bed
with back sticking up towards you
or taped against headboard/wall
angled so back is pointed downish towards bed
Chris Hannemann
@channemann
Sep 09 2015 21:28
I've done float and on hip in bed, both with some successes and failures
Nathan Racklyeft
@loudnate
Sep 09 2015 21:28
@oskarpearson yeah that's an issue in decoding-carelink. If you grab your raw binary data it should be pretty easy to find where the high byte is stored
Chris Hannemann
@channemann
Sep 09 2015 21:28
I don't like leaving it floating because I roll around and end up wrapped in the tubing
Dana Lewis
@danamlewis
Sep 09 2015 21:28
Yea, usually when traveling I do against headboard wall with it dangling or pointing, to mixed success.
Ben West
@bewest
Sep 09 2015 21:28
yeah, out the side is no go
Oskar Pearson
@oskarpearson
Sep 09 2015 21:29
@loudnate thanks for confirming
Ben West
@bewest
Sep 09 2015 21:29
@oskarpearson yeah, if you can get the raw page, it would help
Dana Lewis
@danamlewis
Sep 09 2015 21:30
@bewest interesting letter
Ben West
@bewest
Sep 09 2015 21:30
something like export SERIAL=665455; mm-send-comm.py --init sleep 0; mm-send-comm.py tweak ReadGlucoseHistoryData --save --page 0 should save a ReadGlucoseHistoryData-page-0.data file
you can post the file, or the hexdump, output from xxd ReadGlucoseHistoryData-page-0.data
you might need to change 0 to something else if it's not on the first page
@danamlewis what's your take on it
seems to set some broad/dangerous views, even conflicting potentially with the other things they've said
Dana Lewis
@danamlewis
Sep 09 2015 21:32
just skimmed briefly @bewest but I can see some future implications to us if that's made as a ruling. But again it goes back to distribution or not. Think it more implicates nightscout (does it 'modify the safety of an existing device', placing nightscout in the place as a device manufacturer?) than openaps given our current approaches.
yea, very broad
Ben West
@bewest
Sep 09 2015 21:33
dunno, bakul has already told me he doesn't think nightscout changes existing medical device
that's the whole thing behind "secondary display"
pump vendor could argue that we're changing the "programming" of the pump though
and in fact, that's what's happened, so I'm more concerned for openaps
Dana Lewis
@danamlewis
Sep 09 2015 21:33
he has? interesting. good, but interesting.
perhaps, but we're n=1ing. we're not distributing.
Ben West
@bewest
Sep 09 2015 21:34
so was NS
Dana Lewis
@danamlewis
Sep 09 2015 21:34
with the app, though, it's more distributed-ish now
Ben West
@bewest
Sep 09 2015 21:34
the manuals for pumps all refer to basal schedules and similar config as "programming"
Dana Lewis
@danamlewis
Sep 09 2015 21:35
but there's temp basals for a reason
Ben West
@bewest
Sep 09 2015 21:35
they overloaded the word to mean that changing the basal config is actually "programming" the device and modifying it's usage
Dana Lewis
@danamlewis
Sep 09 2015 21:35
definitely
Chris Hannemann
@channemann
Sep 09 2015 21:35
Yeah, but this line delineates between those cases: "From the discussion, it appears that it would be very difficult for third
parties to fundamentally change the source code of the existing software, but it would be possible
to add or modify software that expands and/or changes the intended uses of the device in ways
that would, under the regulations cited above, require a new premarket notification or PMA be
submitted to FDA before the modified device could be distributed."
Ben West
@bewest
Sep 09 2015 21:35
the assumption in the other feedback is that only the dr makes changes to basals
and temp basals even
there's no consideration that user is already manually making "programming" changes
Dana Lewis
@danamlewis
Sep 09 2015 21:36
@channemann yes, so the difference is whether something is distributed or not.
Ben West
@bewest
Sep 09 2015 21:36
so the perspective here is that openaps and similar are modifying behavior of the device, which is something they are considering illegal
Dana Lewis
@danamlewis
Sep 09 2015 21:36
@bewest i think the word 'programming' was built into diabetes nomenclature before programming everything became a thing doable by everyone
Ben West
@bewest
Sep 09 2015 21:36
sure, maybe
but the effect is very present
Dana Lewis
@danamlewis
Sep 09 2015 21:36
well then every pwd every day is being illegal. that argument doesn't stack up @bewest
Ben West
@bewest
Sep 09 2015 21:36
right, that's kind of my point
or the the point I'd like to make
Dana Lewis
@danamlewis
Sep 09 2015 21:36
if translated literally, yes re: effect being present.
Ben West
@bewest
Sep 09 2015 21:37
read through the other commentary
they literally assume only dr is making changes to pump
user just follows dr instructions
Dana Lewis
@danamlewis
Sep 09 2015 21:37
they're literally.. :::::::closes mouth::::::
@channemann does. I don't ;) (re: doc notes)
Ben West
@bewest
Sep 09 2015 21:37
dunno, I'm a bit worried that the way this is interpreted, openaps is illegal
Dana Lewis
@danamlewis
Sep 09 2015 21:38
It's not illegal
Ben West
@bewest
Sep 09 2015 21:38
distribution is FDA issue
Dana Lewis
@danamlewis
Sep 09 2015 21:38
and its' not regulatable by FDA because it's not being distributed
Ben West
@bewest
Sep 09 2015 21:38
copryright office is regulating the ability to do the modification/reverse engineering
Chris Hannemann
@channemann
Sep 09 2015 21:38
Ha, right
Ben West
@bewest
Sep 09 2015 21:38
FDA seems to be saying it shouldn't be legal
Dana Lewis
@danamlewis
Sep 09 2015 21:38
right, but it's not their jurisdiction
Ben West
@bewest
Sep 09 2015 21:39
right
"not our job, but if you're curious how we feel about it..."
Dana Lewis
@danamlewis
Sep 09 2015 21:39
yea well they get to do that in their letter
Ben West
@bewest
Sep 09 2015 21:39
"we don't really like it"
Dana Lewis
@danamlewis
Sep 09 2015 21:39
will be more interesting to see if anything comes of it
they say all kinds of stuff. you saw the video recently right?
Ben West
@bewest
Sep 09 2015 21:40
apparently not
Dana Lewis
@danamlewis
Sep 09 2015 21:40
(this is the video, linked at the top of my post: http://bit.ly/1LYIjLh)
Ben West
@bewest
Sep 09 2015 21:40
heh, amusing
because people are constantly changing their cars and then killing themselves...
Dana Lewis
@danamlewis
Sep 09 2015 21:41
best analogy ever, huh
Ben West
@bewest
Sep 09 2015 21:41
oh wait... deaths by car are one of top killers... with stock cars!
think he just did our job...
Dana Lewis
@danamlewis
Sep 09 2015 21:41
which job?
Ben West
@bewest
Sep 09 2015 21:41
it's a great analogy
it's one I often use
ok....
he just compared us to anti-vaccers
we really need to make it clear we're using a scientific method, data-based decisions, that's frustrating
Dana Lewis
@danamlewis
Sep 09 2015 21:47
yes
Ben West
@bewest
Sep 09 2015 21:47
but the car analogy is great
people can modify their cars
we don't see an increased risk of death as a result...
we see the market mature
one of those false sylogisms?
Ben West
@bewest
Sep 09 2015 21:53
"40% of literature is wrong"
this guy is fascinating
Ben West
@bewest
Sep 09 2015 21:59
oh, I talked with him on the phone
in march
Dana Lewis
@danamlewis
Sep 09 2015 22:01
which convo was that?
Ben West
@bewest
Sep 09 2015 22:01
was with stayce, alberto, and I
was after I got really really annoyed during follow up with stayce, katherine, beth I think
I detected a shift in their tactics/strategy working with us at the time
where they had previously been helpful, they started being... not forthcoming/helpful
RE xdrip, they asked "why do you need a new algorithm?"
and I wrote back saying that revealed a profound misunderstanding of what this is all about
then stayce + alberto called me to follow up/smooth things over
Ben West
@bewest
Sep 09 2015 22:07
the response immediately after says it all "glad I'm not submitting anything to FDA anytime soon"
Ben West
@bewest
Sep 09 2015 22:31
FDA is really good for growth of the regulatory affairs industry
it's their biggest success
maybe I write a cloud-based dashboard targeting regulatory affairs
probably money in that
Ben West
@bewest
Sep 09 2015 22:59
I dropped bakul patel an emaili
pointed out the video and letter
seems to be inhospitable environment unless you're a monopoly
stuff like this doesn't encourage "the public" to bring material to them
calling patients asking for more data anti-vaccers...
Dana Lewis
@danamlewis
Sep 09 2015 23:09
would love to hear what the response is
Ben West
@bewest
Sep 09 2015 23:26
I'd be shocked to get one ;-)
they don't usually do 1:1 emails
they do the ecopy dance
maybe we consider that, though, just flood them with ecopies