Thoughts about the stock quote database

[Jean-David Beyer]

Rob Browning wrote (in part):
> 
> Dave Peticolas <dave@krondo.com> writes:
> 
> > Rob Browning is going the price database, so he will be posting a
> > requirements/design doc soon.
> 
> Getting started on this now, so it's time to talk :>
> 
> > It may be that gnucash doesn't need to use the price db to store
> > prices from real transactions.
> 
> This, I think, is the first thing we have to resolve.  As I've thought
> about it, though you could put these things into the DB, it would be
> much easier in some ways not to.

Whether or not you store the price database in a formal dbms or not,
what would you use as the primary key for the entries? You cannot use
either the ticker symbols or the company names, since companies change
their names and tickers from time to time, and the tickers are reused,
sometimes very quickly. You need a unique company ID somehow. Also, the
representation of tickers vary from one price server to another.
Berkshire Hathaway Class A, for example, is normally BRK.A, but some
servers require BRKA, BRK/A, BRK:A, BRK.a, BRK A, etc. For many stocks,
these can be the CRSP or the CUSIP numbers, but they are not all that
useful, since I know of no price servers that use anything except the
ticker symbol. They take the ticker symbol and assume the current date
to determine the company wanted. Thus, you cannot look up the old
Anaconda company prices (ticker A, now used by AGILENT TECHNOLOGIES),
nor can you look up the old Chrysler company prices (since C is now used
by CITIGROUP).

-- 
 .~.  Jean-David Beyer           Registered Linux User 85642.
 /V\                             Registered Machine    73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 11:15am up 11 days, 20:03, 2 users, load average: 2.16, 2.14, 2.10