[atlas] Fwd: Re: Proble locations (Changed because of Geo-IP)

Chris,

On Wed, 25 Feb 2015, Chris Amin wrote:

> Dear Martin,
>
> I'm sorry for this, we recently made a change which keeps the location
> of those probes which are geolocated (i.e. *without* a manually set
> address) up to date. Previously such probes were geolocated once, and
> never again. It was never the intention to overwrite manually given
> locations.
>
> Unfortunately, it seems that a number of probes were erroneously marked
> as being automatically geolocated when in fact their owners had
> carefully provided a location. We are now looking at our logs to see how
> many probes might have been affected by this, and whether we can
> automatically revert the location in those cases.

Good. I mainly worry about all the probes where noone noticed the location 
change.

> As for your point about not being able to enter the location, there is a
> text box above the map popup which should allow you to enter a textual
> address. If this isn't working for you, please let me know.

Interesting.. never thought about clicking there, as it already showed the 
*correct* address in it.
Beside of not being able to save, that seems to be an issue with Chrome 
(I've used Chrome 40.0.2214.115 on Mac OSX). Same issue with the 
textfield: It just says "saving..." below and never succeeds.

Trying the same with Firefox worked correctly.

- Martin


> Apologies once again.
>
> Kind regards,
> Chris Amin
> RIPE NCC
>
>
> On Wed, 25 Feb 2015 01:10:30 -0800 (PST), Martin Winter wrote:
>
>> I have actually the address in it.
>> The earlier webpage had a way to enter Lat/Lon for probes (which I can't
>> find anymore).
>>
>> I think a simple Geo-Lokup based on addresses in description would have
>> done a much better job.
>> But a field to enter lat/lon in it and take this as a preference (if
>> entered) without going to overwrite it would be a good start.
>>
>> I mainly worry on whatever happend there in this process may have
>> misplaced a lot of probes to the wrong location.
>> Anyone else noticed that their probe moved to a new (wrong) location?
>>
>> - Martin
>>
>>
>> On Wed, 25 Feb 2015, Grewal, Jas wrote:
>>
>>> Agreed, how about including this critical information in the
>>> description field to get around this, perhaps?
>>>
>>> -----Original Message-----
>>> From: ripe-atlas [mailto:ripe-atlas-bounces at ripe.net] On Behalf Of
>>> Martin Winter
>>> Sent: 25 February 2015 07:29
>>> To: ripe-atlas at ripe.net
>>> Subject: [atlas] Proble locations (Changed because of Geo-IP)
>>>
>>> So I've noticed that (at least some) probes got the location
>>> automatically updated based on Geo IP.
>>> However, this is frequently fundamentially broken. One of my probe
>>> moved from San Jose, CA to somewhere in the middle of Texas. (Broken
>>> Geo-IP location for AT&T DSL users).
>>> (The example I talk about is proble #10770)
>>>
>>> Kind of sad... not sure how many had this happening. Specially as I
>>> took care to put the exact lat/lon in the User interface in the past.
>>> Now I can only reposition it by dragging on a map - and that seems to
>>> be broken as well (at least it fails to save when I try to correct
>>> this probes location on Chrome)
>>>
>>> I wonder how many and why there was a decision to change the location.
>>> I think the location were much better before they got force-changed.
>>>
>>> - Martin Winter
>>>
>>>
>>>
>>>
>>
>>
>>
>>
>>
>
>
>