Op vrijdag 20 januari 2006 00:12, schreef Ognian Pishev: > Cache has a MUMPS background, that is way it is popular among clinicians.
Yes, that is true, Micronetics was taken over by Intersystems, that way they got MSM which is a MUMPS-engine for Windows. Cache still supports the Mumps-programming language, there is where you can see that there must be large portions of MSM-code in it. There is a lot of MUMPS-culture in the code, using %-signs a lot, f.e., and use the word "do" to start something I was on a one-day introduction in Brussels, I saw code, simple code, which create objects, and store them in a database as objects, and retrieve them from the database as objects Cache it is an expensive database-engine. But so is Oracle or SQL-server. They are all expensive. There are a lot of db-engines which are for free, good ones, very good ones. Maybe I understand wrong, but one disadvantage of a OO-db-engine can be data-redundancy. If you avoid data-redundancy it is possible you will be emulating a relational database. For example, if you create a list of address-objects and a list of person-objects, and connect a person to an address from the address-list, then you are playing relational db with an OO-db. Only the way you "talk" to the engine of an OO-db is then is different, but it is very simple nowerdays to hide SQL in classes, so talking to the database can be done without SQL, even if it is a relational database, but of course, there are limits. You can even use programmers without knowledge of SQL, if they exist. So, I don't know if an OO-db has that much of advantage above a relational-db. Often seen in MUMPS, which is hierarchical database (what is in an name), that programmers are not able to take advantage of its features, but start defining relations between globals (how they call there tables), and because MUMPS does not support relations between data, they have to program it, which causes a lot of code and validations to happen. Maybe there is a good reason for the existence of Relational DB's. Ever thought about that? Not only in the old days to save disk space, but in the new days because it is very smart and safe to relate data instead of redundancy. There is one funny particularity about Intersystems. It is a one person-owned-company, not even a family It has no Stock-Exchange notation, and an offer from Microsoft to buy the company was rejected by its owner on other then financial reasons. The owner, Terry Ragon, loves his company, and is commited to MUMPS since the very early days in 1978. So there are no stockholders which decide, but only he. regards Bert Verhees > > Ogi Pishev > ----- Original Message ----- > From: "Bert Verhees" <bert.verhees at rosa.nl> > To: <openehr-technical at openehr.org> > Sent: Friday, January 20, 2006 12:26 AM > Subject: Re: difficulties starting an implementation > > >I wonder, Ian, did you ever looked to Cache (www.intersystems.com) > > It is called Post-relational, this is because it also can understand SQL. > > But it is a kind of OO-database, I don't know if it has all the > > OO-particularities like (multiple)inheritence, or object as field of > > another > > object. > > > > Anyway, a fully functional single user license is free for download. > > > > it is the choice of many large health-systembuilders in the Netherlands > > (Nictiz) and the USA. > > > > regards > > Bert -- Met vriendelijke groet Bert Verhees ROSA Software
