It is not hard to add new instruments in TANGO. Th...
# sciencea
altavir
12/08/2019, 8:43 AMIt is not hard to add new instruments in TANGO. The code base is not quite the state-of-the-art, but it is working properly from the user point of view.
altavir
12/08/2019, 8:49 AM@bjonnh Also since our lab will be doing some device-server-writing work for IAXO, we can also do some device-tweaking work for you. It is my dream to be able to get some money for science by providing services like that (since we are in Russia, it is quite cheap by international standards). But we are also ready to do something like that free of charge in the name of academic cooperation.
b
bjonnh
12/08/2019, 6:17 PMYeah I looked into tango a bit yesterday, not sure it fits what I need here
bjonnh
12/08/2019, 6:20 PMI like the idea, it just seems to be a different realm than what I am trying to achieve. A single protocol for all of these would be really nice though
bjonnh
12/08/2019, 6:21 PMdoes TANGO has to use CORBA?
a
altavir
12/08/2019, 6:23 PMIt is currently being rewritten in zmq. My idea is to separate device API. So we could use lightweight device servers without tango or any other systems.
b
bjonnh
12/08/2019, 6:26 PMyeah that would be great
bjonnh
12/08/2019, 6:26 PM(both aspects)
bjonnh
12/08/2019, 6:27 PMthat's what I tried to do with my system
bjonnh
12/08/2019, 6:27 PMseparate the communication part, the device features etc
bjonnh
12/08/2019, 6:27 PMpretty overkill for just one instrument but if I ever want to add things, I can just decide to communicate through an Ethernet serial adapter instead of direct serial
bjonnh
12/08/2019, 6:27 PMetc
a
altavir
12/08/2019, 6:27 PMThat one I've done. I've even abastracted away the whole communication so I can easily change from TCP to RS-232 or USB
altavir
12/08/2019, 6:28 PMI can port and publish my existing solutions if there is a request for them so I am collecting opinions.
b
bjonnh
12/08/2019, 6:30 PMI wonder how all those things would work together
bjonnh
12/08/2019, 6:30 PMI'm sure everybody is full of opinions on how these things should be done
a
altavir
12/08/2019, 6:31 PMaltavir
12/08/2019, 6:32 PMIn fact, I was working separately and ended up with something very close to TANGO concept. So I am pretty sure that there is some kind of consesus about how it should look
b
bjonnh
12/08/2019, 7:30 PMI'll try to have another look at tango then
a
altavir
12/08/2019, 7:37 PMI do not think that API should be based on Tango. On the contrary, I think we should create API and then add tango connection on top of it
b
bjonnh
12/08/2019, 10:03 PMare you using this dataforge on real things?
a
altavir
12/09/2019, 4:54 AMYep. Old version is used in the whole Troitsk nu-mass daq and analysis. New prototype is in progress, but we already using parts of it in a few projects. A lot of work ahead still.
b
bjonnh
12/15/2019, 12:39 AMLooking a bit deeper into dataforge. Its scope is really large. Things such as fitting, plotting
a
altavir
12/15/2019, 6:51 AM@bjonnh Yep, initially, it was even larger framework: https://bitbucket.org/Altavir/dataforge/src/dev/, but I am gradually rewriting parts of it in multiplatform, leaving the things that could be done by external libraries and improving modularization.
altavir
12/15/2019, 6:52 AMI was an still running the whole-cycle data acquisition, storage and analysis on Troitsk nu-mass on it, but it became too complicated to be easily maintained.
altavir
12/15/2019, 6:53 AMNew version is here: https://github.com/mipt-npm/dataforge-core/tree/dev
5 Views