OpenStreetMap

jfire has commented on the following diary entries

Post When Comment
Running JOSM with OS X Mountain Lion Gatekeeper 8 months ago

My issue was closed with the following comment: "You are using an obsolete JavaApplicationStub which is stuck launching Java SE 6 and is deprecated. You need to rebundle the app with Java 7 (or 8), and resign the app using that."

Running JOSM with OS X Mountain Lion Gatekeeper 8 months ago

I've filed an issue in the Apple bugreporter regarding the incorrect and misleading error message, and encourage anyone with the inclination and an Apple developer account to do the same. (Apple encourages duplicate reports as a means to gauge priority.)

Warning: GPS tracks recorded with an iPhone can be a waste about 1 year ago

Interesting. Here's another person mentioning this issue. There's a suggestion there that the way for developers to switch this behavior off is to set CLActivityType to CLActivityTypeFitness, but I didn't find any confirmation that that works. Do you have a GPS app you could test it with?

Just tried the new iD OSM Editor... over 1 year ago

Yes, as the benchmark shows, there is a large performance difference between Firefox and Chrome on SVG elements, which iD uses heavily. Personally, I find Firefox's performance unsatisfactory. If you're curious about the technical details, the relevant issues are https://bugzilla.mozilla.org/show_bug.cgi?id=837985, https://bugzilla.mozilla.org/show_bug.cgi?id=854765, and https://bugzilla.mozilla.org/show_bug.cgi?id=869505.

Just tried the new iD OSM Editor... over 1 year ago

Here's a test page with no interaction required and a FPS meter built in: http://bl.ocks.org/tmcw/4131834

@aseerel4c26 Have you tried Chrome? I would be interested to hear how it compares to FF on the notebook. (On the Win7 box too.)

Just tried the new iD OSM Editor... over 1 year ago

Hi folks, the first round of performance improvements is live on http://openstreetmap.us/iD/master/ (not on osm.org yet). Let me know if it feels less sluggish. More improvements are in the pipeline.

Some of the changes were fairly invasive, so it's possible I added some new bugs. If you encounter any, please file an issue: https://github.com/systemed/iD/issues

I haven't cracked the problem of sluggish paint performance when panning the map in Firefox. All evidence so far points to it being a problem with Firefox iteself. Here's what we're up against: http://bl.ocks.org/jfirebaugh/raw/5580143/. For me, when dragging on that page, Chrome gets a consistent 60 FPS, Firefox < 1 FPS.

Just tried the new iD OSM Editor... over 1 year ago

@katze_sonne, wow, thanks for the video, that's awesome.

One more thing I should have mentioned -- if you have any browser extensions installed, can you say whether disabling them makes any difference? In Chrome opening iD in an incognito window will usually be sufficient.

I am seeing a few things that can be improved, but I'm also seeing better performance to begin with on my computer. Since everyone here seems to be on either Windows or Linux, and almost all iD developers use Macs, I suspect this may be an OS-dependent issue.

Just tried the new iD OSM Editor... over 1 year ago

@BCNorwich, if the browser crashes it's a bug in Firefox. Not too much iD can do about that. Can you say what area you are trying to edit, and what performance is like in Chrome?

@Tom Chance, thanks, this is helpful. One more question: is there a minimum zoom level at which the performance is decent, and if so, what is it, roughly? (Look in the URL for map=xx.yy -- that's the zoom level.)

Just tried the new iD OSM Editor... over 1 year ago

Ooops, gave the wrong URL above --- try this one.

Just tried the new iD OSM Editor... over 1 year ago

Hi, iD developer here,

We definitely don't expect the kind of poor performance it sounds like you are seeing, and if you're interested in helping improve it, we'd appreciate any kind of detailed technical information you can provide. E.g.:

  • Browser version, operating system, rough technical specs of computer.
  • If you're at all proficient with browser developer tools, any profiling information.
  • Specific areas of the world that are slower to edit than others (if any).

As a shot in the dark as to what could be the problem -- could folks try this version and say whether or not it's any better.

Thanks, John