OpenStreetMap

What is going on with relicensing?

Posted by _sev on 28 May 2012 in English (English)

To me current state and approach to relicensing is an epic fail.

The talks about it started long, long ago, we all know about it, but the actual work on the relicensing bot began in February, with much of the meat added to the code in April, when the work had to actually been already done.

Now we have situation with 14 days with no commits, and in general only 3 people were working on the project.

The main problem is not lack of human power, we (Russian-speaking community) do have it, but absolute lack of visibility. There are no visual traces on what is going on, and messages "5 commits last week" on rebuild mailing list don't really count. No project plan, no requirements, and the worst thing, no one to ask about all of this.

The project is put in a zombie state practically for 2 months now, and looks like a fiasco.

When will it end? Will it ever end? What is the progress? What are the blocking problems? Who drives the effort? Gosh, even tiny things like mapping towns have their own wiki pages with status and leads to the project drivers, but not the whole planet relicensing.

A pity.

Comment from Gnonthgol on 28 May 2012 at 10:00

Everything regarding the license change bot is fully transparant, if there is something you can not find it is because it is not there. You can check out the code at github and look at the mailing list archive, there are nothing more to show.

The reason why there have been little activity is that the work is hard. Even though I have not commited code in a while I am constantly trying to come up with a good algorithm for making the redactions.

There are basically two big problems halting the progress now:

  • Detecting if a change to a tag is 'trivial' or 'significant'. There have been several attempts at this, but there are always some edge conditions that are hard for a computer to detect.
  • Removing changes to way and relation geometries. When a decliner have changed the geometry we need to undo that change without messing up the rest of the object.

If you want to help look at the test cases and add some more cases that is not that hard. If you like to contribute with ideas or code that will be wonderfull. A lot of you do not know ruby (including me) but it is not hard to get a grip on if you know simular languages and are willing to spend some time reading tutorials. If not then contributions in any programing languages or psudocode are welcome. It is a lot easier to translate an algorithm between languages then to come up with a new algorithm.

Hide this comment

Comment from Sanderd17 on 28 May 2012 at 11:17

Gnonthgol, can't you use a bit of crowd-sourcing?

You can pick one specific problem that needs an algorithm, and post it on various social networks. It did already work for code in the Linux kernel: https://plus.google.com/u/0/102150693225130002912/posts/9gntjh57dXt

Hide this comment

Comment from Richard on 28 May 2012 at 13:19

_sev: You're subscribed to the rebuild@ list, I believe. How about offering your help there?

Writing messages accusing others of "epic fail" and "fiasco" is absolutely the wrong way to get something fixed. If you want it to be fixed, offer to help; don't begin by publically calling people out. I'm sure your assistance would be very warmly welcomed.

Hide this comment

Comment from PurpleMustang on 28 May 2012 at 15:16

Personally, I'm enjoying the delay. The longer the better. I'm systematically wacking those multi coloured dots and lines. I'm guessing you could say, I'm using the human algorithm

Hide this comment

Leave a comment

Parsed with Markdown

  • Headings

    # Heading
    ## Subheading

  • Unordered list

    * First item
    * Second item

  • Ordered list

    1. First item
    2. Second item

  • Link

    [Text](URL)
  • Image

    ![Alt text](URL)

Login to leave a comment