Re: [ANNOUNCE] Postgres Success Stories
От | Gene Selkov, Jr. |
---|---|
Тема | Re: [ANNOUNCE] Postgres Success Stories |
Дата | |
Msg-id | 199906100132.UAA12021@antares.mcs.anl.gov обсуждение исходный текст |
Ответ на | Postgres Success Stories (Jeff MacDonald <jeff@hub.org>) |
Список | pgsql-announce |
> Greetings All, > > Do you use PostgreSQL ? I personally don't, but the site I put together at Argonne Lab consistently draws hundreds of users a day, despite its rudimentary interface and almost absent documentation. That's quite a number, considering how specialized the database is: http://wit.mcs.anl.gov/EMP/ This site has also been cloned dozens of times. It's a database with almost no schema that is mainly used for retrieval. it consists of a number of simple index tables; the relations are formed on the fly by queries. It may, however, be of interest to the more general audience because it uses extended data types. > Do you love it ? I sympathize. > Do you want the world to know ? Yes. First, because I owe most of my success to the original and current postgres developers, as well as to the members of the list. Also, because I am worried about the future of my work. The development of custom types and access methods is definitely not in the focus of the popular demand. In the meanwhile, the needs of science are not completely satisfied. We still need a true floating point type that can represent ranges of values, which is required to store the results of measurements in a reasonable way. My own type, http://wit.mcs.anl.gov/EMP/seg-type.html, http://wit.mcs.anl.gov/EMP/seg-type.tgz, is just a successful attempt, and it needs to be further developed and kept in sync with the postgres code, for which I do not have time anymore. I would really love someone to pick it up. I would even argue that such a type is important enough to be one of the core types in postgres. That would allow it to fill the scientific niche as well, where I guess it hasn't so far had any competition at all. --Gene
В списке pgsql-announce по дате отправления: