Hi folks,

Just for some background, there are a few ways that maps can be added to the warper:

Currently - similar to the crop tool - images are added manually from Commons and processed by hand by a user.

There is an admin batch import functionality where images can be imported from a directory of images on the warper server.

There is sysadmin import script where images can be imported again, from a directory.

There is a sysadmin script which can add Ground Control Points to a number of maps from a CSV file.

There is an API where images can be added, control points added and maps georeferenced by a script. 

In instances where the Warper software is in an institution, the Warper has been able to read from an external repository, add the metadata, and import images when requested, and imported images en mass from a database import and finished import when requested.

So there's a few ways that things can be done :-)

Specifics:
> I don't understand why you would want to create a big blue button.  Isn't it easier just to have batch added the identified maps to Warper already?

To help understanding, perhaps think of it as allowing a Commons user interested in a particular map, or group of maps to actively help that map, by allowing them to partake in the whole process.  So I think we are assuming that a User is a Commons user first, not a generic historical maps user. However, the important word from that last sentence is "first" - It's the first step!

In other words, to enable the user to do the whole process is the initial way that it would work but of course it's very possible to change. This is the first incarnation, so to speak!

This touches on the other feature - that of search and discovery. How do we envisage thinking about how users can find all maps within a 20 miles of my location? etc How then would that work with the Commons and with the Warper...These features are all very cool and interesting, but are not effortless to do. So there needs to be something that can easily be done first before the next steps can happen :-)


> 2.  For the program to function, do the images *need* to be physically duplicated over to the Labs server?  Could it not just pull them, when it needs to display them, from a Commons link instead?

The images need to be duplicated over. They are quickly retrieved via URL from the commons link when needed. It could be possible to just store the original images on Commons, it would require a bit of dev work. Nevertheless it would need to store the rectified images somewhere and then there's the question of how to display georectified maps at scale - as cached tiles on a fast server, for example...


Cheers,

Tim

 



On 3 September 2014 13:24, James Heald <j.heald@ucl.ac.uk> wrote:
On 03/09/2014 12:37, Susanna Ånäs wrote:

I think these are the things you propose to develop:
1. Create a mechanism for batch-adding files from Commons to the Warper
2. Create a mechanism for batch-importing georeferencing data to files
already batch-added to Warper.
3. Perhaps make a campaign interface?

All of these require considerable amounts of time from the community and
the developers. We were not prepared for this for this year, but I hope you
will be interested in committing your time to project planning and
resourcing for next year!

Not really.

For (1) and (2) I would simply envisage writing a perl script to automate what would be done by hand.   That ought to be easy enough, reasonably within my capabilities, without requiring much from the rest of the community -- maybe a few tweaks here and there, which is why I'm asking the community what tweaks would be needed.

As for a campaign interface, it can be as simple as a few messages on relevant wikiprojects, plus it would be useful to be able to categorise which maps had not yet been georeferenced, with maybe a progress bar, and a button to offer the user a random not-yet-georeferenced map. Shouldn't be too hard.

So what I'd like to know is, what else is needed?  Or rather, what else would it be nice to have, before rolling this out at scale?

It's a fallacy to think that we can only do so much, because we only have certain resources.  If this is something I can write a couple of bots to achieve, then that's an additional resource to what we had already.

Similarly, if the project excites people and gets them involved, then that too potentially provides additional resources -- eg to translate the georeferencing project page into more languages; and even, potentially, more interest and more resources from people wanting to get in and make more of the rest of the project happen.

Getting something started is infectious.  People start to appear, and want to get involved.

  -- James.


Just to add:

It's better if people can just go and georeference maps, once the facility is available, and the maps have been identified.  I don't understand why you would want to create a big blue button.  Isn't it easier just to have batch added the identified maps to Warper already?

2.  For the program to function, do the images *need* to be physically duplicated over to the Labs server?  Could it not just pull them, when it needs to display them, from a Commons link instead?

  -- J.


Hi all,


Question: what else currently needs to be completed with respect to Map
Warper, before we can launch a mass-upload of files to it, and try to
launch a first mass campaign of geo-referencing ?


So what other things do people think need to be altered on the map warper,
before we
* (a) do a mass upload from Commons to Labs of maps to geo-reference
* (b) launch a campaign for mass-georeferencing the uploads
* (c) use it to display geo-referencing for maps that have already been
geo-referenced ?


I would like to say that our project duration is until the end of this
year, and starting campaigning while we are working on the project will
dramatically reduce our capability to focus on development. But we are
responding to requests of having a user-friendly pipeline available for
georeferencing individual maps by the time of the British Library tagathon.
This will include the following steps:

-- unrelated to georeferencing: --
1. Launch the Template:Map [1], which will include a big, blue button as a
link to the Warper.
2. Use it in a batch upload with GLAMWIKI Toolset [2]

-- related to georeferencing: --
3. Enabling OAuth login from Commons to the Warper has been done.
4. Create a stand-alone big, blue button to add in Wikimedia Commons file
page for maps. Susanna and Jeph work on this.
4. Bot-add this button (a template) to a test set of maps. You might have a
proper candidate for such a set. Eventually we could add this to all maps.
André works on this, but you can help.
5. Create a gadget to manually add the button to any map file. I guess this
goes unassigned.

Additionally, but not before the beginning of October, we will
6. Create a form to transform data from the Information / Artwork / Book
templates into the Map template, that also includes the button, and will
include a locator map for the historical map

During the rest of the year we work to create a new user interface to the
Warper.

Creating methods for easy batch georeffing of maps is not included in our
plan for this year, but we will be happy to include it in a future funding
application. So we decided it will not take precedence over our current
plans.


For reference, as an estimate of the scale of what is on Commons already
(never mind new scans that are being produced), we currently have just over
54,000 unique files in Category:Old_maps and Category:Maps_by_year
together, going down six levels, of which just over 41,000 are in
sub-directories of Category:Old Maps

There are also
* 4890 maps that we have that already have georeferencing from the NYPL
(not in the Old Maps hierarchy)
* 315 Ordnance Survey drawings, with georeferencing from the BL (also not
in the Old Maps hierarchy)
* 1126 maps from the Mechanical Curator collection, of which 107 are
related to images that have georeferencing from the BL, and the remainder
should be geo-referenced in the BL's next phase.


It would be great if someone would put these files into the Old Maps
hierarchy, as that is where the bot could find the maps where the big blue
button will be included.


(The already geo-referenced maps are or will be accessible in
subdirectories of
    Category:Maps with links to external georeferencing
-- the BL are there already; I shall be adding the NYPL ones in a couple
of days, once I have jumped through the required hoops for bot approval).

IMO, it would be nice to get a campaign of geo-referencing underway.

But how close are we to being ready?


I think these are the things you propose to develop:
1. Create a mechanism for batch-adding files from Commons to the Warper
2. Create a mechanism for batch-importing georeferencing data to files
already batch-added to Warper.
3. Perhaps make a campaign interface?

All of these require considerable amounts of time from the community and
the developers. We were not prepared for this for this year, but I hope you
will be interested in committing your time to project planning and
resourcing for next year!

I hope our actions before the beginning of October will be beneficial in
the context of the BL tagathon!

Best,
Susanna



_______________________________________________
Maps-l mailing list
Maps-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/maps-l



_______________________________________________
Maps-l mailing list
Maps-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/maps-l