At present, the known CACHÉ is applied to the hospital information system (i.e., HIS), and it is said that CACHÉ accounts for 70% of the market share in the medical and health industry in Europe and the United States. Donghua Software in China uses the CACHÉ database, and Donghua Software has a market share of about 20% in the domestic hospital market, including West China Hospital of Sichuan University in Beijing.
Caché's features: multidimensional database architecture, post-relational database, built-in development environment, fully object-oriented.
Putting that aside, I really can't think of any reason why the medical system has been running for so many years, other than because of the closure of the medical system.
In the current design, it is full of shortcomings and no bright spots.
Let's talk about the Caché database.
The Caché data is written in M, which is too niche.
The M language was originally born to solve the original medical record management problem of the Massachusetts General Hospital, and in the 60s of the last century, when the R relational database was not born, the M language could indeed bring a lot of advantages to text query and mapping.
The innate problems of the m language lead to the lack of many of the characteristics of high-level languages, ** more obscure than concise.
Data queries rely heavily on indexes.
In an early training session, it took more than 1 week to refactor the indexes for a database that was not very large, which was simply frightening.
Today's Caché deployments are still using a very old master-slave deployment scheme to achieve 24-hour uninterrupted operation.
No matter how you look at it, I don't think it's so reliable.