Page 2 of 2

Re: Proof of Concept:SVN for Mapping & Submissions (For Mappers)

Posted: Wed Jul 22, 2009 9:59 pm
by musang_pulut
Fully support. When to start?

Re: Proof of Concept:SVN for Mapping & Submissions (For Mappers)

Posted: Sat Aug 01, 2009 3:14 am
by chaicka
Juz a quick update.

I have completed the Installation Instructions & Configuration of TortoiseSVN client, posted on the first page of this thread.

Still working on the other procedures.

Note: This is proof-of-concept, and access is limited as it's hosted on a free hosting service. The map tiles hosted are from MFM v1.62.

Re: Proof of Concept:SVN for Mapping & Submissions (For Mappers)

Posted: Sat Aug 01, 2009 3:28 am
by chaicka
For Discussion:

Approach #1:

Use a single main trunk, mappers will lock the map tile intended for editing, perform editing, commit the changes to main trunk, then unlock the map tile.

Pros:
i. Ease of Use.
ii. Small hosting space required.
iii. Low risk of corruption.

Cons:
i. A map tile can only be edited by a single mapper at a time.
ii. If mapper forgets to unlock (release the lock) of map tile, others cannot edit the same map tile until administrator's intervention to unlock it.
iii. Medium-High administrative overhead.

Approach #2:

Use a single main trunk, mappers create branches, perform editing within own branch, commit the changes to own branch, then merge back to main trunk.

Pros:
i. Low administrative overhead.
ii. Map tiles can be edited by multiple mappers at a time.

Cons:
i. Not so easy to use.
ii. Larger hosting space required (dependent on total number of mappers/branches).
iii. Medium risk of corruption.

Re: Proof of Concept:SVN for Mapping & Submissions (For Mappers)

Posted: Sat Aug 01, 2009 9:31 am
by rcmaps
From our experience using SVN hosted free with Google.

It's a real pain to sort out if there is a conflict on map tiles due to multiple mappers editing the same file.

The way GPSMAPEDIT works, any changes moves the relative position of the "lines of code" so a simple small change would render it as a conflict that needs to be resolved.

I would suggest either the locking approach or some way of communication between mappers to ensure only one person is working on a map tile each time.

Re: Proof of Concept:SVN for Mapping & Submissions (For Mappers)

Posted: Sat Aug 01, 2009 10:38 am
by chaicka
rcmaps wrote:From our experience using SVN hosted free with Google.

It's a real pain to sort out if there is a conflict on map tiles due to multiple mappers editing the same file.

The way GPSMAPEDIT works, any changes moves the relative position of the "lines of code" so a simple small change would render it as a conflict that needs to be resolved.

I would suggest either the locking approach or some way of communication between mappers to ensure only one person is working on a map tile each time.


Great sharing. That's what I thought so too. I am in flavor of the locking approach too, since each map tile usually has just a handful of mappers assign to it.

Re: Proof of Concept:SVN for Mapping & Submissions (For Mappers)

Posted: Thu Aug 06, 2009 12:10 am
by chaicka
Status Update:

I have finished the procedures on 'Update' & 'Lock' Map Tile(s) for Editing.

Re: Proof of Concept:SVN for Mapping & Submissions (For Mappers)

Posted: Fri Aug 07, 2009 12:30 pm
by chaicka
2 more free access available to test the PoC (besides AntYong & OngCC)...

Any regional mappers to take both access?

Re: Proof of Concept:SVN for Mapping & Submissions (For Mappers)

Posted: Wed Sep 16, 2009 1:37 am
by chaicka
2 Pilot Testers added. Procedures to 'Commit' edited Map Tile(s) completed.

Re: Proof of Concept:SVN for Mapping & Submissions (For Mappers)

Posted: Mon Sep 21, 2009 11:48 pm
by chaicka
Added my Polygons Submission for 090921 into the 62008888 map tile (besides submitting to OngCC).

Change Log:
chaicka: West (Jurong, Jurong West, Teban Garden) School Polygons Standardization.