Rocket and QM reseller

119 views
Skip to first unread message

John R. Stokka

unread,
Mar 20, 2024, 1:27:58 PMMar 20
to Pick and MultiValue Databases
I'm being told by Rocket that they do not provide dealers/resellers/isvs, whatever you want to call us discounts on QM products or support contracts/renewels because of the "price point".

Are other people in this same situation?

This is ridiculous imho.

They seriously want us to resell this product and make no profit?
What kind of flawed logic is this?

Tracy Raines

unread,
Mar 20, 2024, 3:16:42 PMMar 20
to Pick and MultiValue Databases
I think that Rocket would prefer that you sell a product that you do make money on. I think they only bought QM to get rid of it.

Wols Lists

unread,
Mar 21, 2024, 4:01:03 AMMar 21
to mvd...@googlegroups.com
On 20/03/2024 19:16, Tracy Raines wrote:
> I think that Rocket would prefer that you sell a product that you do
> make money on. I think they only bought QM to get rid of it.

They probably want to consolidate all their Pick assets. Now that
Martin's retired, OpenQM might only have a small team they'd prefer to
deploy elsewhere.

Shame, seeing as it's got the reputation of being one of the best
variants ...

Cheers,
Wol

John Stokka

unread,
Mar 21, 2024, 8:48:47 AMMar 21
to mvd...@googlegroups.com
You've pretty much hit the nail on the head.

--begin rant apologies in advance--

Rocket would clearly prefer that the user base of QM does not expand.
And if we look at the situation objectively.  There is NO competition in the MV market anymore since for all intents and purposes Rocket owns all of the significant viable platforms.

Now that Bain owns the company it's the kiss of death.  Private equity is where things go to be milked for cash flow until they die, especially in an industry with no competitive juices.

All of that being said, the Pick data model has always been a superior model.  It should have long been a native object model in every cloud offering there is.  The only thing that has kept this from happening is a conflict of interest between legacy solutions generating residual cash flow and innovative, disruptive, forward thinking.

I imagine a world where we could develop applications in any language and use the Pick data model for storage.

I imagine subroutines and dictionary items written in Pick/Basic or in C# or any other .NET language for that matter at this point.

I also imagine adapters for javascript/typescript...

Don't even get me started on objects... The Pick data model lends itself to object presentation naturally.

And then there are all of the AI use cases.  We must always remember that under the hood the real value of the Pick data model was performance.  Pure and simple.  The hashing algorithmic storage model is its golden thread.  It's mathematically superior to any other storage model.  This was proven out in the original thesis.  For a relational database, performance is key.  All of the AI use cases are relational data modelling on steroids.  It fascinates me a bit that the folks in the AI world have never stumbled on the Pick data model, or perhaps I'm the ignorant one and they have.

But none of these things will happen inside private equity in a shrinking market.

-- end rant --

In Christ,

John R. Stokka


--
You received this message because you are subscribed to
the "Pick and MultiValue Databases" group.
To post, email to: mvd...@googlegroups.com
To unsubscribe, email to: mvdbms+un...@googlegroups.com
For more options, visit http://groups.google.com/group/mvdbms
---
You received this message because you are subscribed to the Google Groups "Pick and MultiValue Databases" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mvdbms+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/mvdbms/56deeea7-b38b-46a4-90d7-66de87450e61%40youngman.org.uk.
Reply all
Reply to author
Forward
0 new messages