on 3/6/01 10:27 AM, Donna Garverick at [log in to unmask] wrote:

> Richard Gambrell wrote:
>
>> Just about the only possible viable future for the HP e3000 is to
>> stake out the middle ground for business oriented database servers that
>> do Apache, Java and all the necessary up-to-date "stuff", but don't
>> have the cost factor of Oracle.
>
> amen to that!

<snip>

> (boy i hope hp is listening....)
>
> two things:
> - the 'entry cost' for image is too high.  by that i mean, the perception
> remains that in order to have an application on a 3000 that uses image, you
> have to have a custom interface written for the database.  a custom
> interface isn't a bad thing -- you are much more likely to have a well-tuned
> application.  i'll also maintain you're much more likely to find this
> 'design' in legacy applications.  but what about *new* applications?   do i
> train my 20-something programmer how to do image?  (do i even have my
> 20-something programmer working on mpe??)  can i tolerate the performance
> problems of odbc/jdbc?

<snip>

It still seems that HP is pricing to the old terminal based model. When you
start looking at the "necessary up-to-date stuff" a lot of it assumes that
you have CPU cycles to burn. While CPU speed isn't everything, if you went
and told most people that you spent $15,000 on a 110MHz server they would
think you where nuts.

I'm not sure how you balance the pricing for the OS and the speed of the
hardware. It seems like with the latest announcements HP has moved in the
right direction, but it seems like we have a way to go yet.

-Lane