[OSM-talk] Disable Potlatch finally.

Gert Gremmen g.gremmen at cetest.nl
Tue Dec 16 16:08:48 GMT 2008


Anyone shouted BAN POTLATCH!!!11!!!!11!! or similar ???

I politely said without shouting: 

" Disable Potlatch finally. "

No exclamation marks, no capitals.

I still support my plead, but I will
wait with my final judgement  for my POTLATCH experiences with API
0.6...

Although that won't fix the slugginess and associated point and click
errors.

Dave said:
>At least Potlatch doesn't let you delete half of London with three
>keys strokes, or move it three hundred million miles, or upload
>arbitrary OSM files that didn't originate on the server or are 3
>months old,

The small bug-edits are much more annoying,
as no-one notices them. The 3 keystrokes in JOSM
must be made intentionally. Opening this discussion is about 
intentional fraude, not about useability.

Most of POTLATCH problems are clearly caused by:

- not understanding POTLATCH and how it works
- Not understanding OSM at all.

Open data must be open, but not that open
that anyone can run inside and break all the porcelain
without even noticing.
If we won't add a warning label, then
make the entry self explanatory .
(not just a small help window that goes almost unnoticed) 


Patch 1:  (in mappers terms)

Many edit errors are too small to be visible in
Potlatch, lines and dots are FAT and hide small edits.
Suggestion : add further zoom in

Patch #2: (in mappers terms)

When selecting a large OSM area before
clicking edit, it takes quit a time to
load the editor (at least the data)
The area will be too large to edit anyway.
Suggestion: 
Limit the area to be selected (when starting
the editor the first time to say 500x500 meters.

This way the editor responds expectedly
and has a zoom level that fits with the expected
edit zoom level.


Patch #3
permanent column with keystroke help

Patch #4
Before entering the editor a disclaimer
page with concise resources and a warning
against just editing.
Make the play mode the default mode.

Gert Gremmen
-----------------------------------------------------

Openstreetmap.nl  (alias: cetest)


-----Oorspronkelijk bericht-----
Van: talk-bounces at openstreetmap.org
[mailto:talk-bounces at openstreetmap.org] Namens Dave Stubbs
Verzonden: Tuesday, December 16, 2008 4:40 PM
Aan: Steven Le Roux
CC: talk at openstreetmap.org
Onderwerp: Re: [OSM-talk] Disable Potlatch finally.

2008/12/16 Steven Le Roux <steven at le-roux.info>:
> On Tue, Dec 16, 2008 at 1:23 PM, Kenneth Gonsalves <lawgon at au-kbc.org>
wrote:
>> On Tuesday 16 Dec 2008 5:33:18 pm Gert Gremmen wrote:
>>> If a tool has serious problems
>>> it needs someone pointing at it.
>>
>> point out the problems - dont throw away the tool. AFAIK potlatch is
the
>> public face of OSM and the biggest converter of bystanders into
mappers. And
>> for me it is a blessing whenever I have to attend a long and boring
meeting.
>> At the same time, I use josm when mapping an area with a lot of
detail
>> because my 2 GB machine cannot handle potlatch. No potlatch == no new
>> recruits.
>>
>
> I disagree this point. I never used potlatch, because when I tried, my
> couple firefox/debian reminds me it doesn't like flash... which is
> actually unusable.


Works fine here (firefox 3/flash player 9/ubuntu).


> So im not familiar with potlatch...( is each editing frame downloading
> the osm base to see if there is ways or nodes ?  cause there are a lot
> of doubles cause by potlatch, or link untagged...)

There are a lot of API problems that Potlatch exposes because of the
way it works. A lot of these will be fixed in the API's next version.
A lot of the original errors in Potlatch have been fixed by Richard
when people have been nice enough to point them out, rather than
shouting to disable Potlatch.

And BTW, downloading every frame would be more than a little
pointless. Plus how many people do you know who bother to hit download
on JOSM before uploading? -- it doesn't force you to do this, and if
you don't then the results can be a lot of duplicates and broken
links, conflicted ways etc.


>
> this is maybe why I feel close to Gert here... because all the
> pollution, all errors, all unwanted editings are done with potlatch...
> so the shortcut could be easily done. (It's not statistics... it's
> experience... )


"All"? Get real...
At least Potlatch doesn't let you delete half of London with three
keys strokes, or move it three hundred million miles, or upload
arbitrary OSM files that didn't originate on the server or are 3
months old, or.... yes there are errors and annoyances in all the
editors -- this has way more to do with the lax nature of the API than
just the editors in question though.


>
> The fact is... it's true... only patches can correct that. BUT ! for
> the moment... is-it normal to give a that easy access to a tool which
> could cause data loss ? or data pollution...

See above for why we need to ban all OSM editors.
There's very little real data loss as the history is generally intact.
Where it isn't this is the API's fault, not Potlatch's.

>
> Why not prefering to consider other solutions ? Why not downloading
> josm as a java web start app and give in the link the bbox from the
> browser ? has that question been discussed by the past ?


Patches welcome.

But that's no reason to stop people from using a perfectly good tool
because you found an error once which you think was caused by it. Your
JWS JOSM would always be an editing option, not a replacement.


>
>  It's not about disable potlatch, but we have to admin it needs
love...
>

Pity. Means I can't start shouting BAN POTLATCH!!!11!!!!11!! or similar.
:-)

Dave

_______________________________________________
talk mailing list
talk at openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk




More information about the talk mailing list