Forums : Feedback Forum

Ohloh has recently introduced a new feature in the coding experience section.

What open source projects did you use?

The feature is very interesting but it's not clear from the wording if this refers to:

a. libraries used within the project;
b. any kind of open source software that was used (you worked with on a regular basis) during the development of the project;

I think these two questions should be addressed in very different ways:

a. refers to project dependencies and as such it should not (or not only) represented at user-level, but at project level. It'd be great to analyze and represent the connections between different projects as a function of the libraries they use.

b. is indeed user-related, but I don't quite see how interesting it is to suggest that when working on a specific project I've been using - say - an OGG player. On might argue that having a good music player to work with on a regular basis is part of the very experience of coding and - in a sense - can be taken as an open source contribution to the target project. However, this is more interesting to understand how an individual developer works rather than how a project was developed. Also it is not clear if by having experience with an open source project you mean having experience with the software or with its code. I've heavily been using Subversion for the projects I work with but I have no clue of how it is coded. The kind of experience I have with GeSHi (which is one of the libraries used by WikkaWiki) is experience with the code.

Hence my 3 suggestions:

1) Add the ability to indicate project dependencies (a.) at project-level, not user-level. This may allow Ohloh to make some nifty network analysis and to identify projects sharing a large number of libraries.

2) Make stacks more meaningful by allowing some sort of grouping/categorization. This will allow users to describe the tools they regularly work with when developing and distinguish them from tools they use for very different reasons. Also consider that organising stacks by category or kind of use you may produce more interesting ratings (e.g. most popular open source editor, most popular wiki engine, most popular shell etc.) than those currently produced by Ohloh.

3) If you want to keep the current What open source projects did you use? section in the coding experience screen, please make it more precise to make sure that code dependencies and generic software used during development are not confused.

dartar almost 18 years ago

Great post!

First, let me explain my original motivation for the feature. I had 2 goals:

1. Let users track their open source experience. In many cases Ohloh members might not have contributed to Open source projects (like me!) - but we are still interested in showing our involvement with open source.

2. Track who and how open source projects are used. This conflicts somewhat with the stacks feature - but provides deeper context into how a project is used.

Now, regarding the contents of your post:

project dependencies

i agree that dependencies should live at the project level instead of the user level - which is why I don't ask for dependencies in the positions post. The main reason we haven't added this feature at the project level yet is that I believe we may be able to detect dependencies automatically using our source code analysis tools - but that's for another discussion ;-).

I don't quite see how interesting it is to suggest that when working on a specific project...

As I state above, it's just a way for someone to track their experience. Even if you don't know how subversion is coded, you might still be very interesting to me if you know how to administer/manage a subversion repository. I agree it might be useful to know more about how a project was used - although most of the time you can guess how a project was most likely used (subversion? used as an app. GeSHi? coded against the api).

Finally, regarding your suggestions:

Add the ability to indicate project dependencies

Yes - I love this idea. I first need to validate or invalidate that we could do this automatically. Unfortunately this isn't on our immediate horizon so I don't know when it will come about.

Make stacks more meaningful by allowing some sort of grouping/categorization.

I believe this aligns very well with our existing top 2 stack improvement requests:

1. Enable multiple stacks (w/ labels - possibly categories as you mention)
2. Enable personal notes per stack entries (possibly w/ categories too) I suspect we'll get to this within the next month or so.

If you want to keep the current What open source projects did you use? section..

I would like to keep this feature - for reasons that will become more apparent in the next few months. Meanwhile, how would you feel if I added a dropdown that let people choose between:

1. I used this as an application.
2. I coded against its API

Thanks for the thoughtful post dartar.

Jason Allen almost 18 years ago

Hi there,

thanks for your reply.

I believe we may be able to detect dependencies automatically using our source code analysis tools

I had this feeling but didn't dare ask :) Automatically detecting dependencies/bundled libraries sounds like the right way to go, I'm looking forward to this!

how would you feel if I added a dropdown that let people choose between: 1. I used this as an application. 2. I coded against its API

Sounds sensible and I now understand better what your intention was when you introduced this feature.

dartar almost 18 years ago