Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-04 Thread Matt Amos
On Fri, Sep 3, 2010 at 10:14 AM, Nic Roets nro...@gmail.com wrote:
 On Thu, Sep 2, 2010 at 5:20 PM, Ævar Arnfjörð Bjarmason ava...@gmail.com
 wrote:

  You'd want me to spend $12 a day to provide geolocalisation for an
  OSM editor (if you didn't read the thread, I remind you I'm speaking
  of Merkaartor)!!??

 What sort of crazy hosting costs $360 per month (asking Nic Roets)? I

 I said $12 a day for a server that can provide a decent routing service.
 (EC2 High Memory Instance).

 The kernel.org analogy is not really valid. They don't call themselves the
 wiki kernel. When I hack a kernel for an embedded project running on
 specialized hardware, they certainly are interested in my changes.

 By contrast, we will accept users who come in and make a few changes,
 spending as little as 10 minutes.

are you saying that kernel.org can be unfriendly because the amount of
effort put in by each contributor is higher? i don't think that's the
right way to look at it - kernel.org doesn't need to showcase the
capabilities of linux because there are many other providers and
distributions doing this. the people who are using linux, for example
on an android phone, don't need to know about kernel.org, and don't
need to care. the people who are using ubuntu don't need to know about
kernel.org, since ubuntu provides almost all of their needs. in my
opinion it can be the same with osm.org - people can be using the data
(for example, on bing, mapquest or cloudmade) without ever needing to
see the osm.org site. those who are interested enough to want to
contribute can do it directly through the provider, or be directed to
osm.org.

 So we need to make the environment as
 friendly as possible.

i wouldn't disagree with you there, and providing routing on the main
page might make the experience slightly nicer, but it comes with a
cost to OSM in terms of developing, integrating and running the
service which is non-zero. anyone who's interested in getting a
service on the front page has to make it as friendly as possible, not
just to the users, but also to the admins.

my suggestion for getting anything integrated with the main site is to
set up a patched rails port* and your running service on dev.osm.org,
announce it and keep it up to date for a while. when people have been
using it, and it's clear that it's stable and what computational and
admin requirements it has, it has a much higher chance of being
integrated into the main page.

cheers,

matt

* it has to be a rails port and not a static page. either integrating
it into rails port is easy (in which case you can do it - it's easy)
or it's hard (in which case there's a big barrier to getting it
adopted, so you should do it to make it easier to integrate).

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-04 Thread Matt Amos
Kai just showed me your demo running on dev, which is great - sorry I missed
that. The data looks pretty up-to-date, is it being updated weekly or more
often? Is it world-wide?

Cheers,

Matt

On Sep 4, 2010 6:55 PM, Matt Amos zerebub...@gmail.com wrote:

On Fri, Sep 3, 2010 at 10:14 AM, Nic Roets nro...@gmail.com wrote:  On
Thu, Sep 2, 2010 at 5:20 P...
are you saying that kernel.org can be unfriendly because the amount of
effort put in by each contributor is higher? i don't think that's the
right way to look at it - kernel.org doesn't need to showcase the
capabilities of linux because there are many other providers and
distributions doing this. the people who are using linux, for example
on an android phone, don't need to know about kernel.org, and don't
need to care. the people who are using ubuntu don't need to know about
kernel.org, since ubuntu provides almost all of their needs. in my
opinion it can be the same with osm.org - people can be using the data
(for example, on bing, mapquest or cloudmade) without ever needing to
see the osm.org site. those who are interested enough to want to
contribute can do it directly through the provider, or be directed to
osm.org.

 So we need to make the environment as  friendly as possible.
i wouldn't disagree with you there, and providing routing on the main
page might make the experience slightly nicer, but it comes with a
cost to OSM in terms of developing, integrating and running the
service which is non-zero. anyone who's interested in getting a
service on the front page has to make it as friendly as possible, not
just to the users, but also to the admins.

my suggestion for getting anything integrated with the main site is to
set up a patched rails port* and your running service on dev.osm.org,
announce it and keep it up to date for a while. when people have been
using it, and it's clear that it's stable and what computational and
admin requirements it has, it has a much higher chance of being
integrated into the main page.

cheers,

matt

* it has to be a rails port and not a static page. either integrating
it into rails port is easy (in which case you can do it - it's easy)
or it's hard (in which case there's a big barrier to getting it
adopted, so you should do it to make it easier to integrate).
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-04 Thread Serge Wroclawski
 Now knowing that the gazetteer index hasn't been updated in 2 years and was
 considered internally deprecated, it means support for it was implemented in
 Merkaartor (and probably Marble) while the service was already dead but
 still accessible.

I look at this situation as some miscommunication. It's not that the
service was going away in a few weeks, but rather the problem was that
the original service deprecation wasn't widely announced.

I might suggest to OSMers who can make such decisions that services
that are going away have a deprecation warning in them so that it's
clear the service is going away.

Imagine if every Merkaartor user who did a search received  This
service is depricated. Please use Nomimatim. in every search result.
This surely would have triggered a bug report, if nothing else.

- Serge

- Serge

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-04 Thread Nic Roets
On Sat, Sep 4, 2010 at 7:55 PM, Matt Amos zerebub...@gmail.com wrote:
 are you saying that kernel.org can be unfriendly because the amount of
 effort put in by each contributor is higher?

Yes.

Actually kernel.org is very friendly to there audience (Geeks). I'm
thinking of someone like Richard Stallman who does not even surf the
web. lwn.net/Articles/262570/

 people can be using the data
 (for example, on bing, mapquest or cloudmade) without ever needing to
 see the osm.org site.
 those who are interested enough to want to
 contribute can do it directly through the provider,

I'm willing to consider that model.

From Wikipedia's experience we know that people are hesitant to
contribute to anything with a corporate affiliation to it. (Why
should I do anything more than report a bug ? They have so much money
that their people can fix the data )

The other choice would be a second non-profit site. Kind of like
Debian is to kernel.org. The site would try to capture it's own
donations. So it would limit the number of links to osm.org e.g. by
embedding Potlatch. And it would not want to spend money on
development, so it would copy large parts of osm.org. So it will be
just a few steps away from becoming an outright fork... (Banish the
thought).

Kai just showed me your demo running on dev, which is great - sorry I missed 
that. The data looks pretty up-to-date,
 is it being updated weekly or more often? Is it world-wide?

Daily worldwide updates since July 19. The 3 major things that
surfaced / resurfaced were :
1. Gosmore did not handle access=yes correctly, but it's patched now.
2. Motorcars on tracks.
3. Connecting high speed bidirectional roads to the network.

If anyone don't know what 2  3 are about, please ask me, or ask the
routing list.

 * it has to be a rails port and not a static page. ...

Thanks for the guidelines.

Serge wrote
 Imagine if every Merkaartor user who did a search received  This
 service is depricated. Please use Nomimatim.

Good advice.

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-04 Thread Frederik Ramm

Hi,

Nic Roets wrote:

From Wikipedia's experience we know that people are hesitant to
contribute to anything with a corporate affiliation to it. (Why
should I do anything more than report a bug ? They have so much money
that their people can fix the data )


I would love that to be true but the popularity of Google Map Maker 
seems to prove otherwise. On the whole, I don't think people even care.



The other choice would be a second non-profit site. Kind of like
Debian is to kernel.org. The site would try to capture it's own
donations. So it would limit the number of links to osm.org e.g. by
embedding Potlatch. And it would not want to spend money on
development, so it would copy large parts of osm.org. So it will be
just a few steps away from becoming an outright fork... 


I don't see a problem with there being such a site. I'd request that 
they make clear that they are not OSM, and do not own OSM, but I think 
Debian does that pretty well with regard to Linux.


(Ideally, I'd love there to be a handful such sites, just as there's a 
handful of good Linux distributions, each catering to the whims of a 
different audience.)


Bye
Frederik

--
Frederik Ramm  ##  eMail frede...@remote.org  ##  N49°00'09 E008°23'33

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-04 Thread Nic Roets
On Sat, Sep 4, 2010 at 10:53 PM, Frederik Ramm frede...@remote.org wrote:
 Hi,

 Nic Roets wrote:

 From Wikipedia's experience we know that people are hesitant to
 contribute to anything with a corporate affiliation to it. (Why
 should I do anything more than report a bug ? They have so much money
 that their people can fix the data )

 I would love that to be true but the popularity of Google Map Maker seems to
 prove otherwise. On the whole, I don't think people even care.

I think many people are not yet seeing Google as a large, greedy
corporate. Party because the only people who are paying Google any
money are the people who advertise with them.

And if you take the number of MM users as a percentage of the number
of Google Maps users, I think it will be tiny. Only people who are
genuinely annoyed with problems in Google Maps are contributing.

But after doing a few searches like [1], I think we are still ahead of
them. And as we gain market share in the online maps arena, we will
steal MM contributors. I myself was contributing to Google Earth and
Track4Africa* before I started contributing to OSM.

[1] http://www.google.com/search?q=google+map+maker+OR+openstreetmap
* They are currently one of the Google Maps sources for African countries.

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-03 Thread Nic Roets
On Thu, Sep 2, 2010 at 5:20 PM, Ævar Arnfjörð Bjarmason ava...@gmail.comwrote:


  You'd want me to spend $12 a day to provide geolocalisation for an
  OSM editor (if you didn't read the thread, I remind you I'm speaking
  of Merkaartor)!!??

 What sort of crazy hosting costs $360 per month (asking Nic Roets)? I


I said $12 a day for a server that can provide a decent routing service.
(EC2 High Memory Instance).

The kernel.org analogy is not really valid. They don't call themselves the
wiki kernel. When I hack a kernel for an embedded project running on
specialized hardware, they certainly are interested in my changes.

By contrast, we will accept users who come in and make a few changes,
spending as little as 10 minutes. So we need to make the environment as
friendly as possible.
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-02 Thread Chris Browet
On Thu, Sep 2, 2010 at 15:54, Jonathan Bennett 
openstreet...@jonno.cix.co.uk wrote:

  On 01/09/2010 22:33, Chris Browet wrote:

 For myself:
 - Considering the reactions, maybe avoid as much as possible to use
 openstreetmap.org http://openstreetmap.org services, but anyway be sure
 it fails gracefully.

 Chris,

 If you're talking in your role as a maintainer of Merkaartor, then the
 fails gracefully bit is a good idea, but you don't need to avoid using the
 services. Mappers editing the data are the intended audience for the
 services on openstreetmap.org, so Merkaartor is meant to be using them.


Thanks for confirming this. However, the original problem is the same, be it
Merkaartor or a third-party. If the OSM services can disappear in a snap,
it's no real fun for me to develop support for it nor for the end-users to
be confronted to an non-working functionality...

But If everyone agrees that an announcement of ~3 months should be foresee
before a service termination (time to fix and deploy a new version, and some
time for the new version to propagate), it's OK for me.

- Chris -
- Chris -
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-02 Thread Ævar Arnfjörð Bjarmason
On Wed, Sep 1, 2010 at 13:26, Chris Browet c...@semperpax.com wrote:

 If you want to use an OSM service in one of your applications, then I
 would suggest firstly, and most importantly, that the application
 doesn't have a *.openstreetmap.org URL embedded anywhere in the code.

 Come on. Does that mean that we cannot use www.openstreetmap.org/api to
 update the database?

Well, in the context of stable OS releases that distribute programs
like Merkaartor the API is pretty much *guaranteed* to break. An
Ubuntu LTS release is 5 years, Red Hat and Debian have something
similar. That's a lot less time than the entire 0.5 - 0.6 transition
took.

Didn't we have existing packaged programs that used 0.5 break for this
exact reason when we switched to 0.6? If not we're certain to have it
if we ever switch to 0.7, depending on the nature of the API changes.

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-02 Thread Chris Browet
On Thu, Sep 2, 2010 at 16:24, Emilie Laffray emilie.laff...@gmail.comwrote:



 On 2 September 2010 15:20, Chris Browet c...@semperpax.com wrote:

 Thanks for confirming this. However, the original problem is the same, be
 it Merkaartor or a third-party. If the OSM services can disappear in a
 snap, it's no real fun for me to develop support for it nor for the
 end-users to be confronted to an non-working functionality...

 But If everyone agrees that an announcement of ~3 months should be foresee
 before a service termination (time to fix and deploy a new version, and some
 time for the new version to propagate), it's OK for me.


 I think it is safe to say we went a bit fast. Next time, we will be more
 careful in announcing it properly long enough in advance. So you think that
 a period of 3 months is good enough?
 Anyway, we will try to make sure that this doesn't happen in the same way.


Thanks for this.

On a reversed point-of-view, it might also be a good idea to pull the plug
sooner to clinically dead services.
Now knowing that the gazetteer index hasn't been updated in 2 years and was
considered internally deprecated, it means support for it was implemented in
Merkaartor (and probably Marble) while the service was already dead but
still accessible.

Best Regards
- Chris -
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-02 Thread Ævar Arnfjörð Bjarmason
On Thu, Sep 2, 2010 at 15:00, Chris Browet c...@semperpax.com wrote:
 On Thu, Sep 2, 2010 at 16:39, Ævar Arnfjörð Bjarmason ava...@gmail.com
 wrote:

 On Wed, Sep 1, 2010 at 13:26, Chris Browet c...@semperpax.com wrote:
 
  If you want to use an OSM service in one of your applications, then I
  would suggest firstly, and most importantly, that the application
  doesn't have a *.openstreetmap.org URL embedded anywhere in the code.
 
  Come on. Does that mean that we cannot use www.openstreetmap.org/api
  to
  update the database?

 Well, in the context of stable OS releases that distribute programs
 like Merkaartor the API is pretty much *guaranteed* to break. An
 Ubuntu LTS release is 5 years, Red Hat and Debian have something
 similar. That's a lot less time than the entire 0.5 - 0.6 transition
 took.

 Sure, but again, LTS is for business, and I don't see a business using
 Merkaartor ;-)

 And the 0.5 - 0.6 was announced, what, 6 months in advance? With
 discussions and plannification over 1 year...

It really doesn't matter in this context how long it was planned
for. With API 0.6 we didn't have full editor support until relatively
shortly before the deployment (actually only proper support long after
that).

Any distro that took a snapshot of an OSM editor a few months before
the switchover for its stable release would have ended up with an
editor that didn't work anymore.

As long as our API transitions are shorter than major distro releases
(i.e. shorter than 6months - 2 years or so) this sort of thing is
bound to happen.

But there's ways to mitigate that, from an earlier E-Mail you wrote:

 Now, name search on 0.16.x which is going to be included in the
 newest Fedora and Debian won't work :-(

Have you tried to contact upstream about this? This sounds like
something that they'd be willing to patch even in a stable
release. It's a very important feature (in a relatively
unimportant[1]. program) that won't work *at all*. Aren't they willing
to patch that?

 You'd want me to spend $12 a day to provide geolocalisation for an
 OSM editor (if you didn't read the thread, I remind you I'm speaking
 of Merkaartor)!!??

What sort of crazy hosting costs $360 per month (asking Nic Roets)? I
bet you could host basic redirection services on a 512 Linode for
$20/month. Possibly even a whole nominatim instance (I don't know its
resource demands).

Anyway, just pointing the service to your own domain costs next to
nothing. And that seems like a very basic precaution that programs
like Merkaartor should take so they can deal with cases like these.

Have fun.

1. Unimportant in the context of the OS, i.e. it's not like Merkaartor
   is glibc here

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-02 Thread Chris Browet


  Now, name search on 0.16.x which is going to be included in the
  newest Fedora and Debian won't work :-(

 Have you tried to contact upstream about this? This sounds like
 something that they'd be willing to patch even in a stable
 release. It's a very important feature (in a relatively
 unimportant[1]. program) that won't work *at all*. Aren't they willing
 to patch that?


They are very nice people, and I'm sure they would do all they can to sneak
a patch in, but I'm not keen myself in rushing a point release...


  You'd want me to spend $12 a day to provide geolocalisation for an
  OSM editor (if you didn't read the thread, I remind you I'm speaking
  of Merkaartor)!!??

 What sort of crazy hosting costs $360 per month (asking Nic Roets)? I
 bet you could host basic redirection services on a 512 Linode for
 $20/month. Possibly even a whole nominatim instance (I don't know its
 resource demands).

 Anyway, just pointing the service to your own domain costs next to


Err, yes, it would be cheap. But managing/supporting/updating the service
behind it is totally out of scope of Merkaartor.


 nothing. And that seems like a very basic precaution that programs
 like Merkaartor should take so they can deal with cases like these.


I won't discuss this any further, but I still think it is nonsense in the
context of an OSM editor like Merkaartor. Pushing the reasoning, I could as
well have my own copy of the OSM database and push data patches to the main
database...

Mmm... Now that I think about it, it would actually be the thing to do in
the context of the OSM is all about data concept, but it is hardly
practical, resource-wise.

Regards
- Chris -
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-01 Thread Chris Browet


   Is there a low traffic OSM announcement list that provides
 announcements
   like these for people who rely on OpenStreetMap services?
  there is an announce list, but for development announcements like this
  one, this dev list is the best place.

 Well, I just browsed through the announce list and to be honest I think the
 deprecation of a public API really should have been announced there a year
 ago
 if it was already obvious that it would get discontinued.
 Personally I'd say it should be good practice to announce the deprecation
 half
 a year before shutdown of the service on osm-announce.

 I'm afraid I must agree on that (at least 3 months, I'd say, not 2 weeks
for sure)

Merkaartor was using gazetteer  for 1+ year and I only learned of it being
closed down by casually browsing the dev list.
Now, name search on 0.16.x which is going to be included in the newest
Fedora and Debian won't work :-(

As far as I'm concerned, I even didn't know about nominatim, nor the fact
that gazetteer was so behind on updates. I'm afraid there is a communication
channel missing somewhere...

- Chris -
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-01 Thread Chris Browet


 If you want to use an OSM service in one of your applications, then I
 would suggest firstly, and most importantly, that the application
 doesn't have a *.openstreetmap.org URL embedded anywhere in the code.


Come on. Does that mean that we cannot use www.openstreetmap.org/api to
update the database?
That tile.openstreetmap.org can disappear in 2 weeks time?

While I agree it shouldn't be hardcoded, I don't see right now how to make a
difference between a not supported/endorsed/can disappear anytime
openstreetmap.org service and stable ones...
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-01 Thread Torsten Rahn
Hi,

First of all I'd like to say thanks to Grant for providing a solution so 
quickly. :-)

Am Dienstag, 31. August 2010 12:44:54 schrieb Andy Allan:
 Secondly, remember that we're an open-source, volunteer project kind
 of thing, so have appropriate expectations! If the server had stopped
 working and nobody was interested in fixing it, then you'd have had
 zero notice. 

That is pretty clear to me :-)
However as a KDE developer I'm in the Free Software business since more than 
12 years now.
Of course the very same does apply to KDE or Marble (Which is Free Software as 
well):

We do all the development in our spare time. And yes we could just do 
libraries without an API promise and without a promise that any software we 
produce will ever become stable. For stable versions of Linux we could just 
point you to companies who would customize a special stable version from KDE 
trunk and all the other linux software places.

However that is of course not how it works. And I doubt that OSM contributors 
see openstreetmap.org as a nice limited Demo Version of what a true Free 
Mapping portal could look like.

In Free Software projects - inspite of the fact that we do stuff in our spare 
time - we still aim for being able to give guarantees. For the same reason 
that people work on OSM: We want our stuff we work on to be successful.

So in Open Source projects we usually give several kinds of guarantees:
Like 
* API guarantees
* Binary Compatibility guarantees
* The outlook that there will be a next stable version
* The promise that security fixes will go into the next version.

Of course all this only can be made once projects have grown enough so that 
such guarantees can be made. 

I think that OSM is already mature enough that such commitments can be 
introduced. And I think that something like this could easily be communicated 
via properly announced channels.
 OSM already kind of provides such a guarantee implicitely for the tile Urls 
(at least we discussed this topic at the SOTM2008 already). It would be nice 
if this was formally publicized somewhere and extended for other services.

So yes, I fully understand that this is not in place yet. And I fully 
understand why (hey, Marble development processes aren't all rosy either ;-)

But I think there should be an aim to introduce some place where I can look up 
which services/Urls etc. are considered stable and supported and which are 
considered experimental or deprecated. And I think that it should be the aim 
of OSM to have a proper deprecation process. :-)

  On the other hand, if you have someone who's interested
 in running the namefinder service, and have the hardware and hosting
 to do so, then you have more than the OSM project has on all counts
 and that would be great!

Yes, sure, that's why we do already have some map material stored on the KDE 
servers. But you might see why this is a bit problematic for the kinds of 
services and data that openstreetmap.org hosts. 

 The OSM project doesn't run high-availability APIs - there's no
 guarantees, and certainly not any kind of availability guarantee when
 we no longer maintain something that we used to.
 
 If you want to use an OSM service in one of your applications, then I
 would suggest firstly, and most importantly, that the application
 doesn't have a *.openstreetmap.org URL embedded anywhere in the code.
 It's bad practise - see

Yes, that is true. Since the server address is stored inside a xml file it 
shouldn't be too hard to check for a more recent version on usage and use that 
one if available. I had such plans at one point but so far it never got 
implemented.

And thanks a lot for the great work you guys are doing :-)

Best Regards,
Torsten





___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-01 Thread Chris Browet
On Wed, Sep 1, 2010 at 15:49, Patrick Kilian o...@petschge.de wrote:

 Hi,

  If you want to use an OSM service in one of your applications, then I
  would suggest firstly, and most importantly, that the application
  doesn't have a *.openstreetmap.org http://openstreetmap.org URL
  embedded anywhere in the code.
  Come on. Does that mean that we cannot use www.openstreetmap.org/api
  http://www.openstreetmap.org/api to update the database?
 This one might be ok. Iff you make your users sign up for OSM and have
 them read and understand the limits e.g. on acceptable sources.


  That tile.openstreetmap.org http://tile.openstreetmap.org can
  disappear in 2 weeks time?
 Very unlikely. But it might if the hardware catches on fire. And you
 should definitely NOT hardcode that URL into your application. The
 application should talk to your server(s) only. In the start you could
 redirect or proxy the request on your server. If later you app is used
 by 1 million users you can run your own tileserver under that URL. And
 once you reach 2 billion users you can run a load balance there.

 All without changing a single rolled-out copy of your app, just by
 reconfiguring a single point - your server.

 And if the OSM hardware ever burns down you can still feed you
 application from your proxy or even start to render from the last saved
 planet dump.

 While I'd completely agree in a business world, don't forget we are FLOSS
developers working on software in our free time.

We don't (at least I don't, and I doubt Marble has) have a server. OSMF
do.
Just put a big red sticker on the wiki saying ABSOLUTELY DO NOT USE OUR WEB
SERVICES, SETUP YOUR OWN!, then...

- Chris -
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-01 Thread Ian Dees
On Wed, Sep 1, 2010 at 9:14 AM, Torsten Rahn tac...@t-online.de wrote:

 But I think there should be an aim to introduce some place where I can look
 up
 which services/Urls etc. are considered stable and supported and which are
 considered experimental or deprecated. And I think that it should be the
 aim
 of OSM to have a proper deprecation process. :-)


I don't think OSM provides any services that the public should consider
stable in the sense of 99.999% uptime, transaction guaranteed or your
money back. The closest might be the API server, but even that could
possibly go down as has been described before.

OSM is about the data itself, not the services around it. The services
available on the website and wiki are there because someone thought it would
be a nice way to help debug or otherwise improve mapping and map data
quality, not to provide a service to external users like KDE, Microsoft,
Google or anyone for that matter.

The bottom line is: if you want a stable service, run it yourself. That's
why the data is under a relatively free and open license.
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-01 Thread Torsten Rahn

Hi,

 I don't think OSM provides any services that the public should consider
 stable in the sense of 99.999% uptime, transaction guaranteed or your
 money back. The closest might be the API server, but even that could
 possibly go down as has been described before.

I'm NOT talking about a 99,9%-I'll sue you if you don't deliver-guarantee. 

I know that this is not possible to deliver for an open source project.

What I am talking about was _the aim_ and _the strong intention_ of keeping a 
certain API and service up and running for a time that you commit to. I know 
that there can be power outages, lack of manpower or whatever obstacle. That 
is all understandable and natural. 

The only thing that I'm asking for and that I think should be easily 
deliverable is the communicated intention of keeping APIs stable for a 
communicated amount of time. And that there is a communication channel that I 
can subscribe to that informs me about deprecation of APIs soon enough before.

When I release a stable version of Marble I give no guarantee that running 
this version of Marble will not toast your cat. I'm also not giving any 
guarantee that it wouldn't crash. But by releasing a stable version of Marble 
I publically announce: here is something that you can use with least danger of 
crashes. I did my best to prevent crashes and I will surely do my best to keep 
APIs stable. 
That kind of commitment is what I'm asking for. Nothing more. 

 OSM is about the data itself, not the services around it. The services
 available on the website and wiki are there because someone thought it
 would be a nice way to help debug or otherwise improve mapping and map
 data quality, not to provide a service to external users like KDE,
 Microsoft, Google or anyone for that matter.

I could exchange data for Free Software and it's clear that 
OpenStreetMap's position is not at all different from the one of any other 
Open Source project. 

 The bottom line is: if you want a stable service, run it yourself. That's
 why the data is under a relatively free and open license.

With the same attitude I could say: You want a stable KDE or Marble release? 
Check out a development version of the software, compile it and make it stable 
yourself! Works nicely in theory 

Best Regards,

Torsten




___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-01 Thread Nic Roets
On Wed, Sep 1, 2010 at 4:14 PM, Chris Browet c...@semperpax.com wrote:


 While I'd completely agree in a business world, don't forget we are FLOSS
 developers working on software in our free time.

 We don't (at least I don't, and I doubt Marble has) have a server. OSMF
 do.
 Just put a big red sticker on the wiki saying ABSOLUTELY DO NOT USE OUR
 WEB SERVICES, SETUP YOUR OWN!, then...


Just because OSMF, KDE and all the little projects are a non-profit, does
not mean they cannot raise money. Last time OSMF raised money for hardware,
the target was reached in 2 days.

And it makes sense to have a few central servers serving many projects. Not
only does it simplify things like applying patches and performance tuning,
but the planet diffs also has a certain amount of fixed processing to them.
All you do is form an association and pool the donations you receive.

And we know that the service will not be swamped overnight by Google users
or some other large site. Those large commercial companies have high
standards, even for their free products. Our data is not good enough and
they will loose too many users if the site gets overloaded.

And while this discussion has been about searching (geocoding) it also
applies to routing. For $12 a day you can rent a pretty nice server.
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-01 Thread Torsten Rahn
Am Mittwoch, 1. September 2010 17:02:40 schrieb Chris Browet:
 On Wed, Sep 1, 2010 at 16:25, Ian Dees ian.d...@gmail.com wrote:
  On Wed, Sep 1, 2010 at 9:14 AM, Torsten Rahn tac...@t-online.de wrote:
 In our free and open world, I think stable is not about uptime but about
 perennity, i.e., even if the server crashes, it will be there again, be it
 in 1 day, 1 week or 1 month.

  Not even to OSM data editors? For whom, then? Why even announce that
 
 gazetteer was going down...
 Remember the thread is not about the quality of the services provided by
 OSMF, just that the notice was too short...
 
 If some services are just sandboxes and not meant to be used, please make
 it clear, and this is what Torsten means by [a place to] look up which
 services/Urls etc. are considered stable and supported and which are
 considered experimental or deprecated, I think.

Exactly. That's all I'm asking for :-) 

Best Regards,

Torsten

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-01 Thread Chris Browet
On Wed, Sep 1, 2010 at 17:18, Nic Roets nro...@gmail.com wrote:

 On Wed, Sep 1, 2010 at 4:14 PM, Chris Browet c...@semperpax.com wrote:


 While I'd completely agree in a business world, don't forget we are FLOSS
 developers working on software in our free time.

 We don't (at least I don't, and I doubt Marble has) have a server. OSMF
 do.
 Just put a big red sticker on the wiki saying ABSOLUTELY DO NOT USE OUR
 WEB SERVICES, SETUP YOUR OWN!, then...


 Just because OSMF, KDE and all the little projects are a non-profit, does
 not mean they cannot raise money. Last time OSMF raised money for hardware,
 the target was reached in 2 days.

 And it makes sense to have a few central servers serving many projects. Not
 only does it simplify things like applying patches and performance tuning,
 but the planet diffs also has a certain amount of fixed processing to them.
 All you do is form an association and pool the donations you receive.

 And we know that the service will not be swamped overnight by Google users
 or some other large site. Those large commercial companies have high
 standards, even for their free products. Our data is not good enough and
 they will loose too many users if the site gets overloaded.

 And while this discussion has been about searching (geocoding) it also
 applies to routing. For $12 a day you can rent a pretty nice server.


Ok, now I'm in the fouth dimension

You'd want me to spend $12 a day to provide geolocalisation for an OSM
editor (if you didn't read the thread, I remind you I'm speaking of
Merkaartor)!!??

I really don't understand OSM anymore...
When all was needed was Sorry, guys, we'll try to plan our
servers redeployment a bit more in advance and inform you, including service
closure...

I assume the position regarding Nominatim is the same, so I might as well
remove support for it, as it could close anytime. Still wondering why it
exists, then...

- Chris -
___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-09-01 Thread Grant Slater
On 1 September 2010 15:14, Chris Browet c...@semperpax.com wrote:
 While I'd completely agree in a business world, don't forget we are FLOSS
 developers working on software in our free time.
 We don't (at least I don't, and I doubt Marble has) have a server. OSMF
 do.
 Just put a big red sticker on the wiki saying ABSOLUTELY DO NOT USE OUR WEB
 SERVICES, SETUP YOUR OWN!, then...
 - Chris -

Lets calm down a little. :-)

There are 3 reasons why I called for the short notice of shutdown:

1) gazetteer's data hasn't been updated for over 2 years. Internally
it has been considered to be deprecated. Likely we should have
announced this earlier. Oh well.
2) When I checked the usage recently it was extremely low. I wasn't
aware Marble used gazetteer. Marble does not provide a User-Agent when
querying gazetteer. If it had, I would have been less eager.
3) We have other services that are waiting for hardware.

Mistakes have been made... now lets move on.

I'm going to have a hack at creating a namefinder/gazetteer -
nominatim bridge script.

Regards
 Grant

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-08-31 Thread Andy Allan
On Mon, Aug 30, 2010 at 11:46 PM, Torsten Rahn tac...@t-online.de wrote:

 Well, I just browsed through the announce list and to be honest I think the
 deprecation of a public API really should have been announced there a year ago
 if it was already obvious that it would get discontinued.
 Personally I'd say it should be good practice to announce the deprecation half
 a year before shutdown of the service on osm-announce.

The OSM project doesn't run high-availability APIs - there's no
guarantees, and certainly not any kind of availability guarantee when
we no longer maintain something that we used to.

If you want to use an OSM service in one of your applications, then I
would suggest firstly, and most importantly, that the application
doesn't have a *.openstreetmap.org URL embedded anywhere in the code.
It's bad practise - see
http://techfortesco.blogspot.com/2009/10/tesco-finders-towncity-search-not.html
. If your app talks to a domain under your own control, then it gives
you options to rewrite your queries and point them at a different
service.

Secondly, remember that we're an open-source, volunteer project kind
of thing, so have appropriate expectations! If the server had stopped
working and nobody was interested in fixing it, then you'd have had
zero notice. On the other hand, if you have someone who's interested
in running the namefinder service, and have the hardware and hosting
to do so, then you have more than the OSM project has on all counts
and that would be great!

Cheers,
Andy

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-08-31 Thread David Earl

On 31/08/2010 11:44, Andy Allan wrote:

Secondly, remember that we're an open-source, volunteer project kind
of thing, so have appropriate expectations! If the server had stopped
working and nobody was interested in fixing it, then you'd have had
zero notice. On the other hand, if you have someone who's interested
in running the namefinder service, and have the hardware and hosting
to do so, then you have more than the OSM project has on all counts
and that would be great!


I should add that independently of the server on which it was running, 
the namefinder index hadn't been updated in over 2 years. I tried very 
hard last summer and the rate of increase of OSM data defeated me. It 
needs some serious work to make it viable again in an environment where 
the data volume is 10x more than it was when I first wrote it, time 
which I didn't have, and also I didn't see the point when nominatim 
provided essentially that rewrite.


If someone wanted to write a gateway just to convert formats, I imagine 
we could resurrect the gazetteer URL pointing at the Nominatim server. 
Andy's right about embedding URLs, but that's too late for you 
unfortunately.


David

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-08-30 Thread Torsten Rahn
Hi,

Am Montag, 30. August 2010 19:45:20 schrieb Brian Quinion:
 On 30 August 2010 16:33, Torsten Rahn tac...@t-online.de wrote:

  Would there be a way of leaving nominatim support intact via the given
  API just with more limited capabilities (so that it still works with
  reduced quality and taking less hardware resources)?
 
 I think you have confused the services - the nominatim service is not
 being discontinued, rather it is the previous namefinder service that
 is being discontinued.  Namefinder has now been depreciated and not
 updated for just over a year.

Sorry, I was typing nominatim when I should have typed namefinder.

All versions of Marble which get shipped with Linux Distributions are using 
namefinder (and _not_  nominatim). :-(
 
 nominatim is hosted on nominatim.openstreetmap.org
 
 I believe that Marble is already using the nominatim service and as
 such this should not require any changes.

Unfortunately the problem still exists.
The only version that has nominatim support is the Marble version that just 
got released a few days ago. However that one is not significantly deployed 
yet. All Linux distributions are still shipping Marble with the namefinder in 
their most recent version. :-( So the issue is going to persist for about the 
next 6 months.

In that light could you answer my other questions?

Best Regards  Thanks in advance,

Torsten


 
 --
  Brian
 ___
 Marble-devel mailing list
 marble-de...@kde.org
 https://mail.kde.org/mailman/listinfo/marble-devel

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev


Re: [OSM-dev] [Marble-devel] NOTICE: gazetteer.osm.org being retired

2010-08-30 Thread Torsten Rahn

Hi,

Am Montag, 30. August 2010 22:05:59 schrieb Matt Amos:
 nominatim will continue to work and if someone were to write a
 forwarding service which talks namefinder protocol, it could provide a
 stop-gap.

Since we currently pretty much rely on namefinder such a wrapper could 
possibly solve the issue :-)
 
  Is there a low traffic OSM announcement list that provides announcements
  like these for people who rely on OpenStreetMap services?
 there is an announce list, but for development announcements like this
 one, this dev list is the best place.

Well, I just browsed through the announce list and to be honest I think the 
deprecation of a public API really should have been announced there a year ago 
if it was already obvious that it would get discontinued.
Personally I'd say it should be good practice to announce the deprecation half 
a year before shutdown of the service on osm-announce.
 
Best Regards,

Torsten


 alternatively, a commercial service from one of the companies selling
 OSM-based services might be willing to give an enterprise guarantee.
 
 cheers,
 
 matt
 ___
 Marble-devel mailing list
 marble-de...@kde.org
 https://mail.kde.org/mailman/listinfo/marble-devel

___
dev mailing list
dev@openstreetmap.org
http://lists.openstreetmap.org/listinfo/dev