A little list of suggestions

[Atarimaster]

Hello,

over the time, I’ve accumulated a list of things that I’d like to see in Geocart. Here it is.
Nothing is really important, it’s more of a "nice to have" list. The items are listed in no particular order.

– In the "Projection Center" dialogue, I’d like an "Antipodal" button (just like the one from the "Boundaries" dialogue), which of course sets the Antipode of the current projection center.
That’s really not important but occasionally, you might save a few seconds of calculating and typing.


– When you’ve got more than one map in a document, a keystroke to select a different map would be nice. I think the TAB key’d be appropriate. So if one map is selected and you hit the TAB key, Geocart selects the next map, shift+TAB selects the previous map; i.e. the "next" or "previous" map in whatever kind of list Geocart is using internally. Maybe the list that Geocart is already working with when you’re using the "Object -> Bring to Front/Send to Back" etc. menu entries.
If currently no map is selected, hitting the TAB key will select the first map on that list.


– I know that Geocart is meant to prepare maps that’ll be finished using a graphics application. However it’d be nice if some more things could be done directly in Geocart. I especially miss the option to align multiple maps vertically or horizontally (left/right/centered resp. top/bottom/centered). Currently, "align multiple" just aligns maps top and left, but I’d like to say something like: Please arrange the maps left-aligned, leaving the vertical distance as it is.


– I’d like a way to visualize the point(s) of minimum and maximum distortion, both angular and areal, on any given projection.
It’d be fine by me if there’s no corresponding option in Geocart’s GUI but only a new element in the metric line databases, or maybe additional declarations for the isocol element.


– It’d be very helpful if Geocart sanitized the input wherever you can enter lat/lon values before validating.
For example:
• Replacing all commas with dots. If you’re living in a country which *sigh* uses the comma as decimal seperator, you’ll often copy coordinates like 40,751667°N to the clipboard. And then you paste them into Geocart’s input field and hit the OK button, forgetting (again!) that you’ll have to replace the comma first.

• Replacing characters that are often used as degree/minute/second symbols, by the characters that Geocart expects. For example, if you copy coordinates from web sites, instead of a string that Geocart accepts (like 73° 58' 33"), you often get something like:
73° 58′ 33″ W, or
73º 58´ 33˝ W, or even ´´ at the end…
Well, I know, Douglas Adams rightfully said: "A common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools."
Meaning in this case: It doesn’t matter how many characters are being replaced in this way, there’ll most likely still be some weird characters that are sometimes used somewhere that Geocart doesn’t recognize as "character which is to be replaced", but it’d be a great help already if Geocart sanitized the characters I’ve listed above.


– And of course I’d like a few more projections. Granted, I haven’t even fully explored all the projections that are new since v2.3 – but hey, this is the "Wish List" board, so I’ll just list the projections I’d like to have without caring if it seems greedy.
Most notably I’d love the Gringorten, both the square map as well as the variant that is rearranged to give each hemisphere an entire square. I think they’re nice equivalent counterparts to the conformal Peirce quincuncial and Guyou projections.

Other projections:
• Now, we’ve got the generalized Wagner based on Wagner VII/VIII (Thank you!), but the same thing, with less parameters, is also possible for Wagner IX, and even Wagner’s pseudocylindricals (Wagner I to VI). All of them are, in my opinion, less interesting than the generalized Wagner that’s already there, but a generalized Wagner IX would be nice nonetheless. (Personally, I’m not that interested in the pseudocylindricals here…)
• Natural Earth II;
• the Patterson projection;
• the "compromise aspect-adaptive cylindrical projection" by Jenny/Patterson/Savric (which includes the "Compact Miller" that Piotr wished for a while ago);
• and I also like the Bertin1953 projection (see here and here), but I think there’s no inverse formula.


If any of the suggestions seem unclear or should be elaborated on in more detail, feel free to ask.
Kind regards,
Tobias

[daan]

Collating into a terse list:

• Antipode button in Projection Center dialog.
• Keystroke (such as tab) to rotate selection through maps in document, with shift- for reverse. (Do what if more than one map is already selected?)
• Improved alignment behavior, such as aligning only one direction at a time.
• Find and mark minimum/maximum distortion on map, probably as a metric lines enhancement.
• Sanitize text copied from clipboard when pasting as coordinates.
• Gringorton projection, both as square for world and square for hemispheres.
• Generalized Wagner IX.
• Generalized Wagner I–VI.
• Natural Earth II.
• Patterson projection.
• Jenny/Patterson/Šavrič “compromise aspect-adaptive cylindrical projection”.
• Bertin 1953 projection.

— daan

[daan]

Interesting that someone has formulated Bertin’s 1953 projection. A request for that appeared in the forums in 2013, but my exhaustive search for formulas led me to conclude that Bertin never published anything about it. I did acquire the Semiotics of Graphics in that quest, so it wasn’t fruitless.

Also, Geocart doesn’t care if a projection doesn’t have inverse formulas. It uses a generalized inverse if it doesn’t have a directly programmed inverse. That’s generally going to be slower (sometimes much slower) than a formulaic inverse, but it’s extremely robust. Most small-scale projections don’t have inverses reported for them, and even if they do, sometimes I just don’t see the value in writing a bunch of custom code for an obscure projection.

— daan

[daan]

I also like the Bertin1953 projection (see here and here), but I think there’s no inverse formula.

Hm. This is something of a dilemma. I don't think the linked method is a useful approximation for Bertinʼs projection: It deviates blatantly, not just in the oceans as claimed, but also over land.

I hesitate to add just any old bash-it-around-until-we-canʼt-do-better-by-this-technique projection. Approximations can be a useful technique when the original method is not known. For example, Snyder approximated some historical projections, but his approximations truly represented the original intent—probably as closely as there was any intent, even if there were small deviations. I would want something with as much fidelity if I were to claim any association with Bertin. Meanwhile, I donʼt see the merit in adding attempts that, in my opinion, fail, even if they have been made available publicly. Itʼs really easy to formulate new projections these days. I have to draw a line somewhere. If the Attempted Bertin™ projection were to gain some clear currency, I would add it, but I definitely wouldnʼt call it Bertin 1953.

— daan

[Atarimaster]

Collating into a terse list

Thank you, it’s more readable this way.

Keystroke (such as tab) to rotate selection through maps in document, with shift- for reverse. (Do what if more than one map is already selected?)

Hmmm, good question.
As you’ve certainly guessed, I didn’t really think about that before. But now I did. I think the best way (i.e. the least confusing way for the user) would be as follows:

Let’s say you’ve got five maps in the document, and when map #2 is selected and you press TAB, the selection moves upwards to #3, while shift-tab moves to #1.
So, what happens when #2, #3 and #4 are selected and you press the tab key?
Then one of the already selected maps will stay selected, namely #4 – the one with the highest number because you move up.
Shift-tab selects #2.

My first thought was to keep the total number of selected projections (three in the example), but to move up each of them up by one. But I think this gets very confusing: In the example above, hitting tab would select #3, #4 and #5, so apparently only one selection moves. But with six maps in the document and #1, #3 and #5 being selected, all three selection marks jump to a different map.

My second thought was that the selection should move to the next/previous map that currently is not selected, that’d be #5 or #1 respectively. But this can get confusing, too: Imagine that (with five maps in the document again) #2 and #5 are selected. shift-tab then selects #1, and tab selects… #1, too, because the map with the highest number already is selected and Geocart wraps around to #1 again.

We also have to keep in mind that the user probably doesn’t know in advance which map will be selected when he presses the tab key, he’ll most likely hits the key again and again until the map he wants is selected. So I think the best way to make this somewhat comprehensible and predictable is, as I said above, to select one of the maps that already are selected as a starting point. A further advantage is that this also doesn’t break when all maps in the document were selected prior to using tab.

Also, Geocart doesn’t care if a projection doesn’t have inverse formulas.

Alright, I didn’t know that.

Hm. This is something of a dilemma. I don't think the linked method is a useful approximation for Bertinʼs projection: It deviates blatantly, not just in the oceans as claimed, but also over land.

I see.
Previously, I didn’t try to overly the new projection and figure 13 from Sémiologie graphique which they show on the web site. But now I did and even considering that the reproduction of Bertin’s drawing to too small to work with decently, I can clearly see the deviations. Thank you for clearing that up!

Oh well. Meanwhile, I’ve been playing a bit with oblique aspects of other projections and I think I’ve found something that works better for me taste anyway, although I still have to experiment a bit on the exact projection center.

Kind regards,
Tobias

[Atarimaster]

Collating into a terse list:
(…)

• Natural Earth II.
• Patterson projection.

(…)

I just thought of another good reason to include these two projections:
Tom Patterson’s World Political Map.

It’s offered using Natural Earth I, Natural Earth II and Patterson projection. If you want to use Geocart’s excellent raster reprojection on this map engine you already can do that using the Natural Earth I version, but depending on which projection you want to project to, one of the two others might render even better results.

It’s a fine, detailed and professionally designed political map (probably, the best of that kind that you can get for free), so being able to choose the version which is best for a given project would be great.

Kind regards,
Tobias

[Atarimaster]

Here’s another wish:

In the metric line databases, an additional attribute for the isocol element would be helpful:

Code: Select all

<isocol area='2' pivot='0.81831' />

Same functionality as the pivot input field in the Distortion Visualization dialogue, meaning that in this case, the flation value of 0.81831 is treated as 1 and the isocol for 2 gets repositioned accordingly.
Maybe it could also accept a keyword:

Code: Select all

<isocol area='2' pivot='min' />

… which sets the value of the minimum areal inflation that can be found in the regarding projection.

Kind regards,
Tobias

[Atarimaster]

… aaaand one more projection of the „not important, but nice to have“ kind:
The Hölzel projection, a nice modification of Eckert V.
The formula is available in libproj4 and the German Wikipedia („wenn“ = if; „andernfalls“ = else).


I mentioned the Patterson projection above – is there a chance that it will be added to Geocart sometime soon?
Recently I began thinking about a new project, and for that, Patterson would be great. Just asking here, not pressing because I haven’t decided yet if I will carry out that project at all.

[daan]

… aaaand one more projection of the „not important, but nice to have“ kind:
The Hölzel projection, a nice modification of Eckert V.
…
I mentioned the Patterson projection above – is there a chance that it will be added to Geocart sometime soon?

Getting kind of tired of Yet Another Pseudocylindrical Projection.™


I’ll think about it.
— daan

[Atarimaster]

Getting kind of tired of Yet Another Pseudocylindrical Projection.™

… which is comprehensible.
And I have to admit that you can get a result close to Hölzel using Wagner III with standard parallels somewhere around 36°… except for the rounded corners of course (which is why I regard it as “not important”).

As for Patterson, I like its aspect ratio – it nicely fills up the space on an ISO 216 paper size but still leaves enough space for a legend – and because it inflates Europe (where you usually have to put a lot of labels) without inflating the polar areas too much.

Kind regards,
Tobias