[Info-vax] HP wins Oracle Itanium case

[Hein RMS van den Heuvel]

Hmmm, I came here to see whether there was any pertinent info about the topic "HP wins Oracle Itanium case".
I should have known better. The (unfortunately) usually crapperony about Rainbows, proper spelling and other old stuff / water under the bridge from the usual denizens.

But then I stumbled into this RMS related rant from Hoff.
Seems the man has an ax to grind! 
Hoff, and me too (guilty as charged), should have addressed that while we where there when/while we could back in the early 90's. Now we had best make do.

Thank you David Frobble for an admirable reply already.
Just a few items more.


On Tuesday, August 21, 2012 5:21:10 PM UTC-4, Stephen Hoffman wrote:
> On 2012-08-21 20:37:44 +0000, Bob Koehler said:
> o...  and indexed files don't pack old records.  

They do, with severe restrictions. Buckets become associated with a key. A new record with a similar can can re-use space from deleted records in the same bucket.



> o RMS doesn't implement action routines; there are no built-in triggers 

It was frozen in time. This could all have been done if the customer has recognized the need earlier and pounded on, or at least hinted to, product management. 

> o RMS lacks an API and storage for application metadata, beyond 

That gap was more or less adequately filled with CDD / SDL.
Define once, use by many different method.

Later that was further enhanced by 3rd party solutions such as Connx and Attunity with their own dictionaries which allowed further (different) levels of  integration.

> o RMS lacks a data replication API; HBVS and journaling are it.

3rd parties stepped in. 
Connx has had a bucket-comparison based change data capture for a while.
Attunity stepped it up with realtime change data capturing for RMS file complete with target integration (Oracle, SQLserver, Greenplumb, ODBC-generic,...) 


> o Add a key to an existing indexed file.  Or remove it.  Go ahead.  Try  it.  Make my day.

That's the one I personally regret the most for not having fought for a budget to fix that. Dang...


> o No built-in tools to assist with file format and data conversions, 
 beyond CONVERT /FDL and friends.

Well, we had Datatrieve... notably with full domain assignement: new = old.
3rd parties stepped in.
I particularly liked the Vselect tool by EGH. 
Others exists.

> o It's slllllloooooowwwww.

Some beg to differ and consider RMS (indexed) file access nicely fast and predictable albeit with limited (query) functionality.
Yes you can do stupid stuff like millions of duplicates keys in an indexed files or thousanes of file in a directory. But guess what... relational databases and other filesystems have trouble with that too (at slightly higher levels, and with alternate access options).


> [forgot topic]

We sort of had callout / userdefined files  in RMS with the 'SEMANTIC TAGS',
but that was never really hashed out properly (and hashed-keyed files did not happen either ( FABDEF: #define FAB$C_HSH 48  /* hashed */

Cheers,
Hein van den Heuvel (Now with Attunity)