Topic: Isis ATP [Automated Trading Platform] - Discussion - page 20. (Read 45225 times)
lol.
FWIW, it bombs out for me so plenty of bugs still.
-p
Another dead project, ultimately killed due to money interests.
Sad.
-p
I find this somewhat ironic.
We wanted a free open source bot to make money after all.
JP Morgan Chase I would assume.
[snip]
I CAN say that I posted a month or so ago in the OpenPay thread about a conversation I had regarding certain banks taking a long hard look at bitcoin. Lot's of good stuff in there. Just an FYI!
[/snip]
Did any one determine the name of the bank?
tldr
I CAN say that I posted a month or so ago in the OpenPay thread about a conversation I had regarding certain banks taking a long hard look at bitcoin. Lot's of good stuff in there. Just an FYI!
[/snip]
Did any one determine the name of the bank?
tldr
Another dead project, ultimately killed due to money interests.
Sad.
-p
This is a very interesting thread. I'm waiting for something without bugs before I give it a try. Anxiously following here. :-)
Congrats on your gig. That is interesting. Did that corporate contact come about through the Bitcoin community here or did you apply directly and show them what you've been working on? That's really cool, if it was a reach-out through the community.
Good luck (wish I knew Java). :-)
-p
Unfortunately it would be against the terms of my contract to disclose the client or any of their recruiting or operating methods. And I ABSOLUTELY cannot tell you that they contacted me first based on my work on things such as ATP & OpenPay. Therefore I have to remain silent on the subject (for 3 years, gawd I hate NDA's).
I CAN say that I posted a month or so ago in the OpenPay thread about a conversation I had regarding certain banks taking a long hard look at bitcoin. Lot's of good stuff in there. Just an FYI!
Fingers crossed for something by the 8th!
This is a very interesting thread. I'm waiting for something without bugs before I give it a try. Anxiously following here. :-)
Congrats on your gig. That is interesting. Did that corporate contact come about through the Bitcoin community here or did you apply directly and show them what you've been working on? That's really cool, if it was a reach-out through the community.
Good luck (wish I knew Java). :-)
-p
keep up the good work! 
Sounds great. Looking forward to testing.
It's official, I've just been brought on full-time on a project with a major bank (Not JcPenny's but the first letters are JP), to build their next gen trading platform.
Isis-ATP was the catalyst for their bringing me on, that along with my knowledge of Bitcoin. Still I'm not sure what their plans are yet, and I won't be allowed to tell you once I do know what they've got planned, but I've been told I have until Oct 8th to finish Isis-ATP. I'll still be allowed to work on OpenPay beyond that, but they want me to get Isis-ATP out the door before I start with them.
Now we have a hard cutoff date, that means It's crunch time people, so here I go!
(This does mean someone else will have to take over maintainer-ship of Isis-ATP during my tenure with this client (1 year) and during the 2 year non-compete period, so do we have any volunteers?)
any progress updates?
Tons of progress, no updates.
JBoss turned out to be a non-starter, so I migrated the code from EJB to OSGi. The advantage being OSGi modules will run on nearly anything (except JBoss on OpenShift oddly enough) and will allow much of the same code to be deployed to a mobile device meaning all we will need to do is add the UI layer for a mobile. There will also be a stand-alone jar option, it just won't be quite so flexible, but the OSGi modules will be wrapping the standalone jars.
Isis-ATP is being rolled back into the regular OpenPay stack and will just be an OpenPay OSGi module.
Had to refactor an ungodly amount of Bitcoinj & Xchange to make it all work together, but I should have a commit up shortly.
any progress updates?
Here's my report.
Corrected the 2x bidArrow bug in the version prior to arbitrage.
Down about 9% after 18 hours.
Can you post the section of code you fixed? It shouldnt be able to drop like that.
Are these edits being added/incorporated to the source. I haven't downloaded an update recently. Waiting for further development.
Thanks.
Thanks.
Here's my report.
Corrected the 2x bidArrow bug in the version prior to arbitrage.
Down about 9% after 18 hours.
Is this something that would be run locally, or through a central website, with a client? I'm game either way, but now I'm curious as hell.
Edit: This looks like it would/could be run locally, possibly updated on-the-go if you have a server/proper port access. Am I crazy?
Edit: This looks like it would/could be run locally, possibly updated on-the-go if you have a server/proper port access. Am I crazy?
It can be either way. I come from a web services background, so it will be much easier for me to just whip out bunch of servlets and .jsp files and let you admin it through a web-browser.
However the trading platform itself would run as a collection of daemon like modules (yipee no more thread dead-lock issues!). So it's not like you'd need to sit with a browser open all day.
Is anyone opposed to me creating this thing as a jboss module (or collection of them)? Should make deployment much easier, configuration could occur through a web interface and instead of a bunch of .dat files that expire out we could use a database and a message queue to to ensure no ticks get lost.
Thoughts?
JBoss is an Enterprise Application Platform. Basically a container, a set of modules and a collection of tools for building applications in a standardized way.
From your perspective you would simply install JBoss (or go to OpenShift.com) and deploy the .ear file (a collection of .jars) to your JBoss server. JBoss then handles the loading, provisioning of services and overall running of the application. Once it's deployed you'd just use a web interface to configure and administer it just like logging into a website.
Almost forgot to mention, settling on JBoss (or any other EAP), solves a lot of problems for OpenPay in general and I expect to create the reference implementation from it if the Isis-ATP app turns out alright.
As I go through my new design I realize that if I implement it as designed I'm re-inventing the wheel, lots and lots of wheels.
Pretty much all of the components I need to implement this, except the very specific bits related to trading algorithms are already present in products such as Jboss. Additionally OpenShift will let you dev and run for free. I consider that a major plus.
Is anyone opposed to me creating this thing as a jboss module (or collection of them)? Should make deployment much easier, configuration could occur through a web interface and instead of a bunch of .dat files that expire out we could use a database and a message queue to to ensure no ticks get lost.
Thoughts?
Pretty much all of the components I need to implement this, except the very specific bits related to trading algorithms are already present in products such as Jboss. Additionally OpenShift will let you dev and run for free. I consider that a major plus.
Is anyone opposed to me creating this thing as a jboss module (or collection of them)? Should make deployment much easier, configuration could occur through a web interface and instead of a bunch of .dat files that expire out we could use a database and a message queue to to ensure no ticks get lost.
Thoughts?
What's BitMe. Never heard of it. Who is it and and what is your association with them?
BitMe is just another exchange. They charge $0.25USD for ACH withdrawals and my experience has been that they process the ACH requests very quickly. It's irrelevant now that that Bitfloor has come back, since there is now a free ACH path at a place with significant liquidity.
Anyways, crisis has been resolved, thanks!
Jump to: