151 reputation
bio website docsalvage.info
location Tallahassee, FL
age 59
visits member for 2 years, 2 months
seen Dec 31 '14 at 5:25

Unix/Linux software developer in dozens of languages since 1981 with 15 years in Oracle database design and administration.

Projects (most recent first)

  • Developing Zim desktop wiki (for notes) to GtkDialog conversion program to allow automatic creation of checklists for installs, etc. from Zim notes pages. Utilizes bash and 600+ line awk program.
  • Conversion from Tomboy/GNote to Zim for all writing.
    • Zim links to execute or edit scripts with Geany
  • Migrating from Ubuntu Linux to Puppy Linux due to dislike of Unity and Gnome3.
  • Save the world (or at least my sanity) by re-electing President Barak Obama and rescuing congress from the fanatics now running it.
    • 2012-11-06 DONE
      project scope change...
      part 2 improved but not completely doable at this time
  • JavaScript/jQuery-based composing and programming environment inspired by Jeremy Ruston's TiddlyWiki.

answered Reasons for avoiding large ID values
comment Master and Read Only Oracle Database synch is unacceptable
As a large system Oracle DBA who spent 8 months getting Quest SharePlex (probably the gold-standard for fast-as-you-can-get replication) to work reliably, I can assure you that what you're asking for is extremely expensive and very hard to do. One of many scalability ironies... what is simple with a single user system becomes astronomically complex with multiple users.
comment How do I tactfully bring up database shortcomings?
The weekly shrink and reindex sounds like a naive "solution" to a performance problem. The app with the db_owner role again, sounds like a "quick-fix" to satisfy impatient management... the same management you want to talk to about what was done wrong maybe? Think about that.
comment Crow Feet notation and identifying -vs- non-identifying relationships
An Entity-Relationship Diagram (ERD), of which "crows-feet" is just one type of notation, describes only the relationship between entities. The Primary Key (PK) of an entity can normally be inferred from that but there are things you can do in creating a PK that are not clearly reflected in an ERD. By making the Book PK consist of the Author PK plus something else that is not a foreign-key (FK) (you didn't what) you are describing one of those. It would be much more flexible to just use auto-generated sequence numbers for the FKs of all tables.
comment Mysql key_buffer_size already 16GB but some fulltext searches are still slow
This will sound counter-intuitive, but learned from working with large data with high demand a few years ago that buffers can be too big. Searching memory buffers usually does not use the highly-optimized-for-speed techniques employed for disk access as it's unnecessary. Today's memory buffers though are as big as disks of just a few years ago. There's a point of diminishing returns where a smaller buffer will actually be faster.
awarded  Supporter
answered Why are constraints applied in the database rather than the code?
awarded  Teacher
answered In what data type should I store an email address in database?
awarded  Autobiographer