Zest docs browser — gathering feedback

TL;DR: trying to gather some feedback about my idea for another new documentation browser. Mostly prototype screenshots, no release yet.

Writing my previous post about Zest was really exciting for me, it was this early stage of project with lots of fun of prototyping. Now I’m actually losing motivation, so I think another attempt to acquire some feedback could help.

For people not familiar with documentation browsers (like DevDocs, Dash, Zeal), the rough idea is to provide one aggregated place with software development-related documentation, ideally browsable offline.

Here’s what new Zest attempts to bring:

  • Open Source desktop browser for Linux, Windows, and OS X (Zeal can’t do the last one because of an agreement with Dash)
  • Full Text Search with a local index for docs, and for larger datasets like Stack Overflow data dumps. Example simple use case: searching for information about JS promises within Angular docs — the symbols index isn’t enough to find $q.
Full Text Search results from DevDocs and Stack Overflow
  • Access to DevDocs with more benefits of a desktop application, like integration with editors, etc.

What’s new since my last post from Jan 17th?

  • Full Text Search of Stack Overflow dumps now runs in a separate process, which fixes the issue with search causing lags in typing.
  • Full Text Search for DevDocs integrated in the GUI.
  • Prepared some binary packages for Linux and OS X. Unfortunately the build process is not automated yet, but if anyone is interested, I can publish the builds somewhere, and try to do it for Windows too.
  • Implemented more reasonable symbol search strategies. Now exact substring matching and fuzzy matching are working which gives much better results than previous case-sensitive initial implementation.

In general this project was made mostly for fun initially, now I think of streamlining it more, in case people think it is interesting. I realise some development choices I made can make it more difficult:

  1. I basically prototyped it quickly without much care for organising the code well.
  2. I’ve chosen the most ‘fun’ language for me personally (ClojureScript), as opposed to something most practical.
  3. Instead of reusing as much as possible from DevDocs, I’ve decided to make my own GUI. This is a partly a side effect of 2 above.

However, given some feedback/support, I think I’d be willing to work on 1, on making 2 a more viable choice, and hopefully justify 2 by creating some reasonable new UX, with more focus on desktop.

Unfortunately my previous post didn’t receive any comments, so I’m not sure if it really makes sense to put time into it, given it lost some of the ‘fun’ factor already. I’d love to build some community around the project, but it seems hard to find any places to make it possible. Perhaps it’s in too early stage, but still I think it’s worth trying.

You can reach me via comments, or email my firstname dot lastname at gmail dot com.