Hi all,
I am trying to get to grips with julia and to achieve that I have set myself the goal of putting together a pure julia approach to a python app I wrote. Because of the way I wrote the python app I missed out on a LOT of the functionality of python because I just coded on the fly with no regard to design. Once it was working I moved on. The ONLY thing I spent anytime doing properly was pandas and that was time well spent.
The issue I had, note the past tense, was that I had an interface to Interactive brokers api via a WONDERFUL gateway api called insync ewald ( the author) had done such a great job that I didn’t have to consider the manner in which it worked. It just did and I could get on with what was important to me knowing the data would flow. THEN my ignorance began to show when I tried to hook it up to julia. I got excellent help using PyCall and could get many things working, when my mental model was set straight, but NOT the ability to tap into the data stream. I tried all sorts of hacks ( writing out to csv files and reading them back into julia), trying to tap into the STDOUT buffer, using ZMQ, sockets et al. All seemed to be ridiculous to me AND detracted from what I really wanted to do, move to pure julia but still using interactive brokers api . You can actually use other languages ( look at the top of the docs and select one) but none of them were julia. HO hum…
THEN along came @nilshg with a great suggestion, why not use ib api in julia. I would NEVER have found this as I don’t use github so hat’s off to Nils and discourse!!!
and now my question, this is an unregistered package and knowing IB things change, I am concerned that it might not work in the future and I might not be able to use or fix it. So what are people’s experiences of this type of concern, am I tilting at a windmill here? I had the same concern when I started to use pandas until I looked deeper into Wes. I can’t do that with jib.jl because there is only one email registered and no one is answering that one.
It seems to be actively developed, so why not open an issue asking for it to be registered?
At the same time, it is MIT licensed, so essentially you can do anything with it, which means that if it is abandoned someone could fork it, register it and assume the development over (all that assuming that the original authors are not interested anymore, if they are the best is always to join efforts). That may sound harsh, but that is FOSS.
hi there @lmiq thanks for the insights, please remember I don’t use github so bit hazy on the whole approach. I thank zarquon that some excellent people even got working on Jib.jl thus giving me:
1 hope I can use pure julia
2 a platform to learn
3 a grounding in github should I need it.
part of the difficulty is that I can’t really ask questions about Jib.jl as I don’t know how to, I assume that there are few people on discourse using it and I would feel bad about asking them to get up to speed just to help me. I try to not bother people until I have done some gruntwork.
It is perfectly fine to ask them here, the worst that can happen is not getting an answer.
To open an issue, just create a github account (it is free, you don’t even have to be “real” in any meaningful sense there). Then go to Issues · lbilli/Jib.jl · GitHub and create a “New issue”. I think asking that package to be registered is a reasonable demand, since it has already a reasonable user-base.
thank @lmiq I’ve marked this as the solution as it’s given me excellent insight into the github approach. I haven’t really gotten into it as I tend to develop my “code” on my own and dig myself out
One last question, you said it had a reasonable user base but I can’t see how you know that? I see 8 people watching.
It has 31 stars and has been forked 8 times. That is not irrelevant. (The stars are the most common measure of popularity of a project, I would say that anything with more than 10 stars is probably something reasonably meaningful).
