hi

Hi @alvin-c-shih - there was a brief mention on the All Community call last week that the KDB group were interested in contribution towards a new UI? I think someone mentioned that KDB Studio is a NetBeans app and that you wanted something more modern?

We might be able to help with this - could you perhaps elaborate a little?

(I was going to attend the call today but it is a bank holiday In the UK!)

@ColinEberhardt Yes, the motivation is that the current editor backend is very dated so features commonly found in modern editors are missing. So either it could be migrated to the latest version of the NetBeans codebase or any other modern Java-based UI framework. This would affect the majority of the code in the project as it was basically forked from NetBeans at some point, and the KDB-specific parts are relatively small.

Thanks for the info @gyorokpeter - does it have to be a Java-UI? We do very little Java-based UI, the vast majority is JavaScript / TypeScript, using React or Angular.

Main points:

1. kdb+ client API. Would have to be able to port this without significant runtime performance penalty (say under 2x): https://github.com/finos/kdb-studio/blob/master/src/main/kx/c.java
2. Pluggable authentication mechanism. Each environment might use something slightly different: https://github.com/finos/kdb-studio/blob/master/src/main/studio/core/IAuthenticationMechanism.java
3. Export results to Excel in a way that doesn't trigger warnings.

As long as these requirements are met (plus being reasonably modern, of course), almost anything goes. :-)

Additionally the aim of this project is to be a desktop client, rather than another browser tab. This is the very reason I'm still using this program instead of Kx Analyst which is available with the enterprise KDB license (or Kx Developer if you don't have that). So the result of doing such a drastic change would be a completely new project and not a new version of KDB Studio.

Thanks @gyorokpeter and @alvin-c-shih - I certainly don't want to push you down a technology route that you are not happy with. For a bit of context, I am the CTO of Scott Logic, a UK-based technology consultancy. I'm always on the look-out for interesting open source projects that our consultancy can potentially work on between client work. When I heard about this one in the All Community call it sounded interesting.

We're pretty flexible about which technology we use, but it does have to be something that reflects the type of work we typically do for a client.

With that in mind, is it worth us doing a little bit of a feasibility study to see whether using JavaScript / React / Electron (making it a desktop app), is a viable option - meeting the performance, auth and Excel integration requirements mentioned above?

Have you considered doing it as a vscode plugin? Thst way, you can use JavaScript and node to develop, and you still have a desktop app. Many similar things, let it be jupiter notebooks or python support, is already there, so it could be a nice addition. Also, distribution is provided free by the marketplace.

Hi @psmulovics - completely agree, VSCode is a very good option for this sort of environment. Have you seen this?

https://github.com/jshinonome/vscode-q

@alvin-c-shih @gyorokpeter - before we invest any time in exploring this further (i.e. looking at technology options, planning / sizing the project), I do want to confirm that you think it is worthwhile? We don't want to invest time looking into an option that isn't right for this project!