Do not hardcode styles

[tmcw]

Not sure if there's a ticket for this one, but this is basically priority #1 that marks OpenLayers as opposed to any design thinking. Stuff like the panzoombar has no excuse to be inline-styled and needs the styling to be ripped out into stylesheets.

[crschmidt]

How is this not a duplicate of openlayers/openlayers#35 ?

[tmcw]

It is; search didn't turn that one up. Closing.

[crschmidt]

Also, if you don't like the controls, don't use them. The right solution almost certainly isn't to 'port' these ancient controls forward, it's almost certainly to throw them away -- by putting them into deprecated.js -- and write something at least plausibly sane. These controls were never good, have not been properly maintained, and more modern browsers and a fresh set of eyes could almost certainly rewrite them in significantly less code -- and unlike most 'old' code, I don't think that there's any plausible excuse that PanZoomBar continues to have any value because of hairy bug fixes. The only reason it hasn't been rewritten is because nobody cares. It's not going to happen until someone who needs a panzoombar (few apps want one) that looks better sits down and writes it.

[tmcw]

People use the OpenLayers default controls; this guy Barack has them on his site, broken images and all: http://www.whitehouse.gov/change

There are two solutions:

1. Actually remove the controls
2. Improve the controls so they feel like 2012 code, not 2006

I like #1 and you do too; but given that this ticket is 6 months old and the code is 6 years old, so I severely doubt this will ever happen.

As far as #2, you should care; see above. This is a serious, user facing issue that makes the experience of using OpenLayers worse. The library needs to be improved, in concrete ways like this, not just ballooned up with new features.

[jorix]

For my #2 (something must have OpenLayers, not only trust that exists GoeExt)

But it is difficult to propose changes from outside, eventhough I am aware of the problem

[crschmidt]

Misconfiguring your application can cause your application to fail regardless of how you use it. If you remove your image directory from a CSS file, you'll get broken images. Pretending that this is a result of code being bad is silly: the same problem would happen with the PanPanel control, which uses CSS and doesn't have hardcoded images.

Moving the PanZoom, LayerSwitcher, etc. controls through the deprecation path seems completely reasonable to me, but other OpenLayers contributors can comment.

I don't see any evidence that this is a high priority issue for OpenLayers. Having long-provided the ability to use CSS-styled controls for the typical default control (including the one used in the URL you referenced), people don't use them: if people don't use them, then it's not clear that this is worth bothering to learn something new for their use cases. With that evidence in hand, I still don't see evidence that someone other than you thinks this is crucial.

[tmcw]

Moving the PanZoom, LayerSwitcher, etc. controls through the deprecation path seems completely reasonable to me, but other OpenLayers contributors can comment.

See other issues on deprecation for background. The general impression for me is that the lack of enthusiasm around deprecating features and modularizing the library has been the nail in the coffin. Deprecating even the most obvious wastes, like utility functions used once, has been a pain, and getting a new API to break these 2005 paradigms has been a years-long struggle.

This is a high priority; people switch because of this and it completely prevents any design-minded organization from using the toolkit.

[jorix]

@crschmidt

I don't see any evidence that this is a high priority issue for OpenLayers...

also not a priority for me.

... people don't use them:

EG see http://sitmun.diba.cat/sitmun_gca/?mun_ine=08006 an example of the regional administration of the entire area of ​​Barcelona that uses PanZoomBar.

And PanZoomBar have bugs not related to css: drag to the end and drop, or drag off the map and drop ...

[tmcw]

What is a high priority issue for OpenLayers?

[crschmidt]

My first priority has always been: Continuing to work for the existing users of OpenLayers.

I can't actually speak in any significant way on behalf of OpenLayers -- I'm no longer an active participant so I'm not really in a position to comment on what OpenLayers devs want. I apologize if I made it sound that way.

You seem to be treating "People switch from OpenLayers to other libraries" as a problem. I don't understand that. Why is it a problem that people use other libraries that fit their needs better? OpenLayers isn't going to solve every problem ever. If it's bad for developers, in a way that affects you (but you can't fix), then perhaps other tools are the right tool for the job. shrug

[jorix]

@tmcw
At the individual level I'm sorry that is not a priority for me. It's just reality: requires time and have already 6 Pull Requests for review.