Re: TIDE, Railhead, and Oracle

  • From: "Niall Litchfield" <niall.litchfield@xxxxxxxxx>
  • To: david.barbour1@xxxxxxxxx
  • Date: Mon, 8 Sep 2008 22:40:33 +0100

Take a look at 
http://democrats.science.house.gov/Media/File/Commdocs/Staff_Memo_toBM_terror_watch_8.21.08.pdf

Hmm, so staff and contractors added tables and views without reference
to the data model, tables are undocumented and people don't know what
specific attributes of specific objects mean. The tables are unindexed
(maybe something else is meant). A single business event "searching a
suspects pockets" gives rise to entries in 23 seperate logical
entities. Clearly this is all the fault of SQL (one has to wonder
since the project started life in XML if this gloss doesn't come from
the 'architects').

It may be that this isn't a bunch of disparate agencies throwing their
existing systems into a new central bucket because they were told they
must, but I wouldn't bet against it. If I'm right the criticism of SQL

SELECT TARGET FROM SUSPECTS WHERE SOUNDEX(KNOWN_AS) LIKE 'OSAMA%';

as opposed to XML and a app server language.Seems somewhat off target,
and indeed a misreading of the memo.

Niall




On Mon, Sep 8, 2008 at 5:11 PM, David Barbour <david.barbour1@xxxxxxxxx> wrote:
> I've no idea why this could possibly be allowed to happen.  Coming back
> through Immigration from a recent scuba diving trip, I marveled at the total
> inefficiency of the passport control process.  Why (I thought) for US
> passport holders couldn't we just walk up to a kiosk, swipe or scan our
> passports and have some type of visual recognition software match our mug
> with the photo on file?
>
> I guess if they can't manage 500,000 names, they couldn't possibly do
> anything this horribly complex.
>
> On Sun, Sep 7, 2008 at 9:09 AM, Bill Ferguson <wbfergus@xxxxxxxxx> wrote:
>>
>> I sure hope nobody on this list is working on these projects. This was
>> just forwarded to me by my Project Chief.
>>
>> Apparently the House Science and Technology Committee finds SQL lacking.
>> Laying off 800 contractors? - be careful who you hire. Good thing it only
>> cost half a billion dollars.
>> It's the 4th story down here.
>> http://news.cnet.com/the-iconoclast/?categoryId=9756918
>>
>> Sounds like it was originally written in Oracle/XML. Railhead team wanted
>> to convert it into Oracle proper. It cannot do Boolean keyword searches?
>>
>> http://mobile.eweek.com/device/html_article.php?id=1&CALL_URL=http%3A%2F%2Fw
>>
>> ww.eweek.com%2Fc%2Fa%2FGovernment-IT%2FKey-US-Terrorist-Database-Program-Mir
>> ed-in-Controversy%2F
>>
>> Poor oversite to blame in this article. 463 tables, 295 of which are
>> undocumented.
>> http://blogs.zdnet.com/projectfailures/?p=995
>>
>> <rant on>
>> If anybody on here is on either of these two projects, use Oracle Text
>> for the searches. Send me an email and I'll send you a Word document I
>> created for my Project Chief that shows how I created a little (3200
>> line) PL/SQL program to loop through each table gathering all the info
>> for each 'parent' record and write this into a CLOB field in XML
>> format. I also have triggers on all my tables so every time a record
>> is added or modified, the XML/CLOB field gets ceated/updated, and then
>> all searches get performed against that XML/CLOB field with Oracle
>> Text. After setup (which took several days on my database), the
>> searches are done in a matter of seconds. I have it designed for a
>> simple Google-type search, or you can even specify which fields (XML
>> 'tags') to search within.
>>
>> It ain't that dang hard! Why can't I get one of those jobs? Oh yeah,
>> I'm already in the government. Maybe I can use these articles to
>> justify a promotion?
>>
>> </rant off>
>>
>> --
>> -- Bill Ferguson
>> --
>> //www.freelists.org/webpage/oracle-l
>>
>>
>
>



-- 
Niall Litchfield
Oracle DBA
http://www.orawin.info
--
//www.freelists.org/webpage/oracle-l

Other related posts:

  1. Home
  2. oracle-l
  3. Archive
  4. 09-2008