Charts II: NGA - 2700 Charts

[wdkester]

If we are going to lock charts, we need an "editor" who is responsible for checking the charts, locking them, and then unlocking them and/or repairing them if found to faulty. Perhaps we already have this, perhaps not. Perhaps the responsibility could be distributed; a small number of people who lock and unlock charts and a larger number of people who have the responsibility to check the charts and recomend the locking/unlocking.

Whether the charts are locked or not I would like for us to start with something simple like a facility for chart comments or enhancements, arrainged one discussion per chart, most recent comment first. Here I see things like: 1. a link to an overlay with NTM's, 2. user experiences like: "I have plotted tracks on this chart and the chart stinks", "My, previously user, waypoints fit perfectly on this chart, "There is a new cruise ship dock in the middle of the harbor". 3. Armchair observations "Google Earth is inconsistent with the shoreline on the East side of the harbor"."Google Earth seems to show a datum shift of about 100 meters on this chart".

Even the "simple discussion" per chart idea might need a moderator to eliminate the extraneous or malicious comments.

Start simple and build. We still have a lot of work to do on insets.

[HappySeagull]

"Chart Git-Wiki"





(cropped from File:Gldt.svg - Wikipedia, the free encyclopedia)

[HappySeagull]

Lock em when they're all done...Then,maybe there could be something like checkboxes added.....x number of users can say NoGood and why and next it (only)gets unlocked by a (volunteer) "webmaster".Subsequently,it's fixed by any logged-in user who wants to apply the agreed fixes,test it and relock it.... so as not to put all this on the "webmaster"...if we stay within the parameters we've used so far(just what's available on the jpegchart),thiswon't be too much,I bet....the rest of it,wikis and versions could be hosted here?or anywhere that can provide the space and people to man it...

[cagney]

Why the rush to release?
There is still a lot to do.
Many of the charts marked as doable in stage 1, needs special PLYs, for example. Not to talk about all those time consuming inserts and charts just consisting of plans.

The charts we have now, are best labeled as alpha testing.
Next step could be to release a beta test for a popular area, when all the charts, for the area, are complete. The point of this would be to test our efforts among more users. Let's get some wider feedback, before we talk about how to organize and release the charts.

What do you guys thing about the quality of the produced charts.
Do you realize that the are reduced to 40% of the original? Is this the right balance between size and quality?

To the question about how these charts are used. Every chart has this text embedded in the kap headers.

!This chart originates from
!http://www.nauticalcharts.noaa.gov/mcd/OnLineViewer.html]NOAAOn-Line Chart Viewer[/url]
!DO NOT USE FOR NAVIGATION
!Use official, full scale nautical charts for real-world navigation.
!These are available from authorized nautical chart sales agents.
!Screen captures of the charts available here do NOT fulfill chart
!carriage requirements for regulated commercial vessels under
!Titles 33 and 46 of the Code of Federal Regulations.

When we release, this text will be repeated in release notes etc.
There will probably be a note similar to the one OpenCPN forces all users to approve when used the first time, as well.

Thomas

[HappySeagull]

on this chart,18419 vs the real one 341501 (canada) comparisons image size are 4640x6400 vs 7305x9739
it's not possible to zoom them the same (though the upcoming openCPNGL version might do this!) I have some other comparisons versus old maptech hdr that look good too.

I still have my doubts..sure ,the caution note is very plain in OpenCPN but I wonder if it shows in other Nav apps?Ozieplorer opens em but says nothing...and, while the downloading person may "agree" and "approve",after that,won't the kaps themselves show up all over the place,similar to mp3s,copied hdd to hdd,cd to cd?oh well.

[bcn]

Originally Posted by cagney

Why the rush to release?
There is still a lot to do.
Many of the charts marked as doable in stage 1, needs special PLYs, for example. Not to talk about all those time consuming inserts and charts just consisting of plans.

The charts we have now, are best labeled as alpha testing.
Next step could be to release a beta test for a popular area, when all the charts, for the area, are complete. The point of this would be to test our efforts among more users. Let's get some wider feedback, before we talk about how to organize and release the charts.

What do you guys thing about the quality of the produced charts.
Do you realize that the are reduced to 40% of the original? Is this the right balance between size and quality?


Thomas

Hi Thomas,

first if somebody wants to use the .kap that are generated he can do so right now. Downloading some 1700 files one by one is a little "pita", but everybody is free to do so.

For me question is how to evaluate a chart and to fix what we expect to call it "ready for sailing".
- complete (no extrusions, indents or those covered)
- the grid fits?
- also in the center and 1/3 from the corners?
- how does it compare to other charts (if it is different, that does not mean it is worse though)?
- are the datum data correct?
- does it run in different programs (which are the benchmarks to check against?)?
- .....?

I do agree with you that there is a big amount of work ahead still (insets etc.), but that does not mean that the charts that are ok (see above) must been treated as an "alpha" or immature result - or?
For me for instance the lower resolution does not make them useless.

The process of releasing should happen step by step. And perhaps it is open ended.

Hubert

[HappySeagull]

...well,at least they don't look exactly like "real charts",so anybody with a little experience is going to recognize these right away...
Caveat emptor and the price is right!

[z.stadler]

It seems like some members feel uncomfortable with a public release. On the other hand, it seems like we do need a specific next goal to keep the momentum of the community.

I'd like to suggest a few possible directions:

1. Complete the "Rectangular, not doable in Phase 1 because the corners are not visible on cutouts" and "Rectangular, not doable in Phase 1 because of the 90 degrees skew"
2. Clarify the approval process, approve the "Rectangular, doable in Phase 1" charts, and create a beta release of this subset.
3. Extend the OpenCPN.info site to enter PLYs for "Rectangular, not doable in Phase 1 because of special PLYs needed", approved them, and add them to the next beta release
4. Extend the OpenCPN.info site to calibrate the insets in "Rectangular, partly doable in Phase 1 - With rectangular insets", approved them, and add them to the next beta release
5. Combine 3 and 4 above to support the insets of "Rectangular, partly doable in Phase 1 - With non-rectangular insets", approved them, and add them to the next beta release
6. Discuss and agree on the requirements for a public release

Your thoughts?

[bcn]

Originally Posted by z.stadler

It seems like some members feel uncomfortable with a public release. On the other hand, it seems like we do need a specific next goal to keep the momentum of the community.

I'd like to suggest a few possible directions:

1. Complete the "Rectangular, not doable in Phase 1 because the corners are not visible on cutouts" and "Rectangular, not doable in Phase 1 because of the 90 degrees skew"
2. Clarify the approval process, approve the "Rectangular, doable in Phase 1" charts, and create a beta release of this subset.
3. Extend the OpenCPN.info site to enter PLYs for "Rectangular, not doable in Phase 1 because of special PLYs needed", approved them, and add them to the next beta release
4. Extend the OpenCPN.info site to calibrate the insets in "Rectangular, partly doable in Phase 1 - With rectangular insets", approved them, and add them to the next beta release
5. Combine 3 and 4 above to support the insets of "Rectangular, partly doable in Phase 1 - With non-rectangular insets", approved them, and add them to the next beta release
6. Discuss and agree on the requirements for a public release

Your thoughts?

It seems to me that Nohal is already working on the points 1, 3 und 4.

The main remaining issue is the release and maintenance process and how/where to host the (organized, bundled) finished files.

And I would like to understand better the doubts about the quality, usefulness against to release already the (to be) finished material.

So 50/50% d'accord - but asking for a debate.

Hubert

[HappySeagull]

Originally Posted by bcn

It seems to me that Nohal is already working on the points 1, 3 und 4.

The main remaining issue is the release and maintenance process and how/where to host the (organized, bundled) finished files.

And I would like to understand better the doubts about the quality, usefulness against to release already the (to be) finished material.

So 50/50% d'accord - but asking for a debate.

Hubert

...hard to get a debate.Everyone roughly agrees.Good Ideas and drift like pipesmoke while we hang out in the mess waiting to be piped into action....
I don't have doubts about quality.I know they are supplements.Nohal is out testing em where he is...
If Moby Dick is out there,it's probably live navigators that'll see him.

Meanwhile,I'd like to hear from Cagney,recri,Paulr and and anyone else that's scrutinising as to a best method to double-check charts that others could follow.Keep track of the ones you've done and any comments.

[nohal]

As someone immediately discovered, I'm back
All the broken charts rotated, corners extended etc. This is probably the last round on these problems - there will really be just a couple left and I will do them myself without further notice.
Other progress:

• As you may have noticed, there's a new link to the GPX file defining the chart polygon next to each chart now - just ignore them for now, they will be used in the process of defining the non-rectangular polygons for the charts when I finish all the needed scripts (now really soon )
• The backend is now dumping all the data to https://github.com/ngacharts/data daily, so we have another backup of all the work done

More later

Pavel

[z.stadler]

43301 - SW Corner was set out of bounds, need to be reset on the server side.

[recri]

On the question of release issues. All charts are inherently imperfect representations of reality. Even the most perfect chart can be invalidated in a few days by a big rainstorm that moves all those carefully identified shoals around, or by a big storm that ends the useful existence of clearly identified aids to navigation. And many charts aren't that perfect. Some are truly scary.

So these chart images, copyright without restriction, were of varying degrees of perfection when we started. Our work to georeference and correct them makes our versions of the charts into derived works to which we hold the rights.

I think we write a Creative Commons license that spells out our reservations in detail and strongly encourages the user not to be an idiot. Someone might want to ask Creative Commons and see what legal assistance they can give us on this. It's an interesting licensing issue because the consequences of misuse are potentially so dire.

Downloading the charts should require acknowledging the terms of the license. We include the license, or a link to it, in every chart we release. One of the terms of the license is that we forbid removing the license on penalty of losing our permission to freely use the charts.

Then we need OpenCPN, and every other chart plotter, to be encouraged to make the user aware the licenses and/or terms of use of the charts that it has open.

RNC Downloads spells out the terms of use of the NOAA RNC charts. You're in violation if you haven't checked for and downloaded updates in the past week, and you're a class of ship allowed to use the charts for navigation, and you're using the charts to navigate. OpenCPN, and other chart plotters, should be making you aware of that, not assuming that you've got it under control.

-- rec --

[Viking Sailor]

Guys,

Paper charts have been passed around from person to person forever. These charts are no different, not only will they be downloaded, they will also be passed around on DVDs. Thus, any terms of use check boxes on a download site are going to be largely irrelevant. IMHO, all that is needed is a "not for navigation" watermark. That is all I have ever seen on any non-original or historical chart. This is the "common practice" for distributing this class of charts. So, please, don't drive yourself crazy thinking you need to do more then what is the common practice.

And, remember the open source mantra -- release early, release often.

[recri]

On the question of a git-structured chart repository.

This is more of an open question than an immediate goal. I think the advantages are clear: you want the most up-to-date charts the community has generated, you want to remember more up-to-date corrections if you've verified them in person, you want to share those personal discoveries with the community, and we want to incorporate all information that makes the charts more useful. So how can you do that? Can you do that for everybody? Or does it only work if you're moderately obsessed with chart accuracy?

At the level of KAP formatted charts, the first part is somewhat easy. You maintain the list of charts you want to keep, and you write a script using rsync (or some other time based synchronizer) that fetches newer versions of your charts when they become available. It requires rsync services (or some other time based sync server) at the server hosting the charts, and that you run the update scripts when you have bandwidth to do the work.

But given that we're constructing the NGA chart set from raw images and a bunch of geometrical notes, one could save a lot of bandwidth by only downloading the updated geometrical notes and regenerating the chart with the previously downloaded image. The geometrical annotations used to turn the raw images into KAP's, that's the database that Nohal is maintaining, could be easily maintained in git with a complete history of modifications, the ability to add or correct notes in your personal branch, push your modifications back to the central repository for merging, and pull updates from the central repository.

You could be generating your own raw images, too, by only downloading the required tiles from NOAA at the resolution you wanted, applying a rotation if necessary, and assembling them locally. The history of updates to the raw images at NOAA would be helpful, since updated chart images might not align with the georeferencing of old images.

So, I guess you'd want a chart manager, a program which fetched the catalog(s) and let you pick the regions you needed to be kept up to date, the resolution you wanted to maintain for the different scales of chart and different regions, and made it so.

Feeding back corrections is a more complicated problem. Most people could care less, and no user interface will change their minds. But if you're sufficiently paranoid and meticulous, you'll probably put up with almost any interface that does the job.