Slashdot Mirror

Check out Intersystems Cache. http://www.intersys.com/ It successfully merges SQL and objects in the same database. Very nice. (Disclaimer: I work there).

As for MUMPS on Linux...yes, you can. Aside from as-yet incomplete projects like Generic Universal MUMPS (gump) that are actually somewhat useful, but not yet ready for "prime-time" development, you could take a look at Cache from Intersystems (the folks that currently own DSM). It's a "full-featured" commercial MUMPS that runs like stink on Linux. There's a trial version (no fees required), and a single-user license (which enables full networking) won't break the bank. I've been playing with it for a while now, and can report that yes, it is MUMPS...for whatever that's worth. It's also supposed to really rock for database-driven web serving, but I haven't as yet been able to substantiate those claims, as at work we're still stuck in the 60s (using none of the newer APIs).

As for being more powerful than BASIC...well...maybe the case could be made that the pattern matching and string handling routines are better ('cause they are!), but as an all-around language, it stinks. MUMPS' claim to fame was based in its integrated database--not in its flexibility. For example, some versions limit individual strings to a max length of 256 bytes...ugh.

Regardless, it is possible to make money in MUMPS.

Hmmm...speaking of my resume, looks like I've programmed professionally in FORTRAN77, COBOL, SYNON (an AS/400-based 4GL), and MUMPS. Man, I feel like a dinosaur!

In addition to the links above, most of the big database systems have active Linux ports. Any Oracle, Sybase, Informix or DB2, InterSystems, Poet, or Versant customer is a potential Linux customer.

One of Epic's biggest competitors is IDX, of Burlington, VT. Their product is also based on M.

This is not cheap stuff. If you're concerned about the price of Windows, you probably can't afford it. You'll pay the most for the medical software and support contracts, then for the underlying DBMS (e.g., M, Oracle, SQL Server), then for the hardware (e.g., 8-way SPARC with gobs of RAID storage), then for the OS. Don't forget the salary of the local staff to support the system.

If you're serious about this, you might consider talking to a salesperson at a major DBMS vendor and asking for a referral to one of their VARs (value-added resellers).

Big database software installations, especially medical software, are a pain. At Epic, a six-month install was considered impossibly short. The sales process alone usually took that long, involving a five-hundred-page RFP (request for proposals) from the customer, a thousand-page response from Epic, and a meeting of the lawyers to sign the contract.

If the secret to individual success is "underpromise and overdeliver," the secret to corporate success is the exact opposite. Sales will say, "of course we can do that." After you've signed the contract, Development and Support will say, "are you nuts?" A month later, after $400 / hr. of custom programming, it will sort of work. Bring a systems analyst to the sales meetings--someone who can kick the tires, ask questions, and understand the answers.

I've felt for a long time that database software is a good candidate for free software development, because many of the customers already employ a large technical staff, including developers. If you stick to the standards (SQL, etc.), you'll have a stable base and a large community. Most of all, nobody cares about the product as much as the customer. Epic, IDX, and InterSystems are for-profit companies that sell proprietary solutions in competition with other such companies. They care about the product insofar as you'll buy theirs instead of someone else's.

It's not Open Source, but you can try Cache` from Intersystems It can talk SQL, and stored procedures are written in M. It's used extensively in the medical industry (where I work). They license by concurrent user, not sure about pricing though.

I'm not a database expert, but I've done a fair amount of healthcare IT consulting over the years, and run across Intersystems' Caché database quite often.

Most people aren't familiar with Caché from Intersystems, but if you're building a very large, sparsely populated transactional database (like an electronic medical record, for instance), it's at least an order of magnitude faster than Oracle, even after doing unnatural things with Oracle. (This info came directly from a major EMR vendor that benchmarked Cache and all the usual suspects and found nothing in the same league for this application. According to his tests, Caché was 30x faster than Oracle out of the box - Intersystems' website claims Caché is 20x faster than RDBMS competitors.)

The technology is quite interesting in some regards: It is very mature, having its roots in the old MUMPS or "M" system (about as old as Unix), but has been updated quite nicely as time has passed to provide quite capable SQL and object-oriented interfaces, even though it is neither an RDBMS or an OODB internally.

Another nice feature is its distributed caching protocol, which allows you to build a logically huge database server piecewise from a number of distributed servers. A number of large installations, including Boston Children's Hospital (IIRC), use this for their EMR - last I heard, BCH's database was spread over 100-150 servers. This would seem to fit nicely with the Linux way of doing things.

It runs on Linux and a free download is even available on thier website, or they'll send you a CD for the asking if you're one of the unfortunates stuck on the end of a phone line.

You can find Intersystems at http://www.intersys.com, not exactly the world's most intuitive URL, so I thought I should include it here...

Intersytems has a product called Cache that is a post-relational database (i.e. heirarcharial) It is very fast. The now have a Linux version. (Was originally VMS)