Re: Foreign key in Oracle Sql

Hi DA,

On Mon, 24 Jan 2005 08:51:38 -0800, DA Morgan wrote:

>Hugo Kornelis wrote:
>
>> Seems we have a terminology mismatch. Tens of thousands of employees is
>> large in my perception. Hundreds to thousands is middle-sized; one- or
>> two-digit personnel numbers is small.
>
>That's what I thought. To me this is the definition:
>http://www.sba.gov/size/sizetable2002.pdf

Thanks for the pointer. I'm more acquainted with the Dutch definition. I couldn't find a reference on the web, though.

>>>But even if you aren't watching ... MS is ... and Linux is the flavour
>>>of the decade.
>>
>> Which decade?
>
>2001-2010. And if that is not true the paranoia here in Redmond
>Washington, with respect to Linux, would dictate that it will be
>very soon. Balmer is down in Brazil begging the country on bended
>knee not to walk away from Windows and the PRC (China) and other
>countries, such as Germany, are walking the same road.
>
>The entire planet does not revolve around the US

That's what the rest of the world has been trying to tell you for years already - I'm glad that at least some of you are now also realizing this!

>and our influence
>is decreasing as each day passes.
>
>> http://www.emarketer.com/Article.aspx?1003196 - in 2004, 1.3% of PCs
>> worldwide run Linux as opposed to 96.2% running various versions of
>> Windows.
>
>And what were the numbers 5 years ago?

These are much harder to find on the 'net. Most sites with old numbers are either actualized with newer numbers, or no longer up, or they are listed only on Google's page 12,345 or so. I did manage to find a few tidbits:

• http://www.muq.org/~cynbe/rants/lastdino.htm This page, written jan 1999, claims Linux has reached 2.5% of the desktop market. If that were correct, Linux would have lost market share over the last five years. But I don't take this site too serious - the writer also claims that the most likely growth extrapolation is a doubling of the market share each six months, which would put Linux at 100% somewhere in 2002 (unfortunately, the author fails to explain what another doubling of market share, from 100% to 200% would indicate - each 'puter running TWO copies of Linux???)
• http://news.com.com/2100-1001-236732.html?legacy=cnet This page has a more serious appearance. It lists Linux market share for server OS at 16% in 1998 and 25% in 1999. I'm not sure how to relate these numbers to the numbers I quoted from other sources in my previous message (projected 28% of worldwide server shipments and redeployments by 2008; 16% of enterprises expecting over half of the company servers to run Linux in 2005).

>The numbers you quote include desktop machines and home machines which
>are irrelevant in the context of this usenet group: comp.database.theory.
>
>No one working with serious databases is using them at home or on their
>desktop. Serious databases are in the realm of servers and server
>clusters where the numbers are substantially different. A look at the
>decisions being made by Sun with respect to Solaris, HP with respect
>to HP/UX and IBM with respect to AIX vs Linux indicate my statement is
>valid in this context.

I'm well aware of that. That's why I also included the projected share of server shipments and redeployments by 2008 (28% - source: IDC, as quoted by eMarketer) and the outcome of a survey by Evans Data among enterprises about the expected number of Linux-running servers (only 16% expect more than half of the servers to run Linux).

>>>>They are concerned with running their business, not with learning
>>>>computers, so they LIKE a pretty and intuitive GUI. They are already used
>>>>to the Windows look-and-feel. They'd have a hard time converting to any
>>>>other OS.
>>>
>>>Exactly what is the difference between the Windows GUI and the RedHat
>>>Linux GUI?
>>
>> (snip)
>>
>> Why then did *you* write:
>>
>>>>>Linux rather than Windows. Not have the GUI overhead.
>
>Because in the context of a server one doesn't type STARTX at the
>command line and there is no GUI (X-Windows) overhead. That is not
>an option with Windows.

Granted.

>>>>The CPU spends most of it's time twiddling thumbs.
>>>
>>>Not in the database world which is the topic of this discussion. The CPU
>>>often is pegged at 100%.
>>
>> To be more specific, the topic of discussion is databases being used as
>> backend for an ERP package that's being deployed at a small to medium
>> sized business - up to a few thousnd employees in my perception.
>
>I'll grant small but I still don't see that as mid-size.
>
>> If the CPU is pegged at 100% in that scenarion, then they aither use a
>> very inefficient database as backend, or they have some terribly
>> ineffective application logic.
>
>Or they do things that are math intensive such as aggregations which is
>the heart of pretty much any reporting process.
>
>> Within the context of databases used as backend to ERP packages, more
>> functionality of the database is irrelevant.
>
>Nonsense. That is absolute nonsense. If one were to agree with your
>statement one would have to also agree with the statement that much
>of the functionality of SAP, PeopleSoft, Siebel, Oracle CRM, and Onyx
>are unnecessary: And they are the industry leaders.

Maybe I didn't explain myself too well. The context of the discussion was: small (I'll stick to your terminology) companies running ERP with MSDE as back end - if they outgrow MSDE and need to move to a larger DBMS, SQL Server Standard edition is the perfect choice. Those companies won't use the claimed extra functionality of Oracle. If the ERP package needed that functionality, it wouldn't run on MSDE. If it runs on MSDE, it uses functionality present in MS SQL Server and it won't suddenly start to use extra functionality if Oracle is used instead.

Best, Hugo

-- 

(Remove _NO_ and _SPAM_ to get my e-mail address)