[Loadstone] Possibility to ignore Cellid...

Shawn Kirkpatrick shawn at loadstone-gps.com
Fri Oct 9 00:29:00 BST 2009

Could you clarify a couple things for me? First, I'm not familliar with the 
term bts, what does this mean? What are these sector numbers you're talking 
A database of cell sites would be very useful, I wonder how much data we can 
actually get and for what countries?
Your change sounds simple but there's a couple questions. Is the cellid 
value ever actually 0? Wouldn't there have to be some result processing as 
well to see if the query returned an exact match as well as a result with a 
cellid of 0?
I've been thinking Loadstone's cell handling needs quite a bit of work. 
Mainly a separate database for cell sites so they're not so connected to the 
point data.

On Fri, 9 Oct 2009, Grzegorz Zlotowicz wrote:

> Hi,
> sometime ago I found the site where people collects all possible bts informations unfortunately for Poland only, but there are plenty of such sites for other countries too...
> I wrote a converter of data which mentioned site is providing, and I'm testing its' usability for Loadstone users at the moment.
> Having prepared such huge database of bts lacs and cids for your gsm network, that it is nearly complete is very useful, letting you know where aproximately are you at the moment without need to switch on the gps receiver, especially great in the bus or train...
> There is also a bit of fun knowing on which street and building is the unit, that your phone is connected to...
> Only one problem i found is the cellid value - Loadstone stores both Lac and Cid of actual place, it's great in many cases, because the last digit of cid is the sector identifier, and the bts sectors are in most cases more preciselly identifing the place that only location area code...
> The problems are, that when you're connecting to the same bts but with umts not gsm, the sector number is different for the same place...
> Some bts also can have more than one sector covering the same spot, to make the load of network less...
> But, returning to subject - in mentioned bts database all sectors of bts are described in the same way - city, street and additional informations - sometimes not all sectors of the bts are entered into the database...
> The main disadvantage of this is, that the cell table is overloaded by the same information appearing many times, and if some sector of known bts isn't described, and unfortunately your phone connects to this sector, LS behaves like you are on the area of unknown bts...
> I'd like to suggest a very small modification of Loadstone, letting it ignore the cellid value, but also keeping current functionality unchanged...
> And, most important, implementing it is a matter of only one line of code!
> The cellid value is the number greater than zero, so let's assume that if cellid in the Loadstone cell table is 0, it is a description of whole bts, not only one of cells.
> It seems to be extremely simple to implement, but i'm not sure if the Symbian db engine would let construct query with parentheses, and if it would make the searching process slower...
> So, my suggestion is to change the line number 142 in the waypoint.cpp file (src of version 0.72):
> Originally it is:
> (start of query not quoted here) locationareacode=%u and cellid=%u");
> After the modification:
> (start of query not quoted here) locationareacode=%u and (cellid=%u or cellid=0)");
> I tested this query in the dbshell program, and it seems to work.
> What do you think about implementing such change?
> This simple modification would be useful not only for users of Loadstone from Poland, also from an other countries where people provide a comprechensive listings of machine-readable bts stations.
> Greetings, Greg.

More information about the Loadstone mailing list