MultiValue – the next “hot new” thing

I would have no problem evangelizing the MultiValue data model as the “hot new” database alternative to the worn relational model. The problem is that none of the MV DBMS vendors seem interested in doing this.

This quote was posted to the U2 forum today. I’m not trying to say that the comment is bad, just using it as a common example of how even people who love the platform can be its worst enemies.

I find it amusing how ‘recent research shows…’ that column based storage is quicker then your standard relational database. I found this company VectorWise(http://www.vectorwise.com) because they are
partnering up with Pentaho, a BI suite that I am learning. I see these new databases popping up here and there talking about a “New” way of storing data (column-based, tables-in-tables, etc.). We may not be ahead of the curve as far as front-end processing, but MV is the only way to go for storage!!!

That’s a nice comment but it’s both complimentary and (erroneously) insulting to the platform.

As to “We may not be ahead of the curve as far as front-end processing” : Why do people who spend their lives with this platform still think and say things like that? There are any number of great tools for development of GUI, Web Services, and interfaces with platforms like Microsoft Office apps, and telephones, and all kinds of other software. This is how I spend all of my time, what do you mean “we may not be ahead of the curve”? A public statement like that by people who actually use the software is just another nail in the market coffin. Where is the VAR of someone who makes a statement like that, to re-affirm on a daily basis that the platform is indeed fully capable? Where is the upline DBMS vendor to make sure that “Value-Add” resellers are representing their products in a manner that reflects positively on the products and market?

As to “MV is the only way to go for storage” : It’s nice to sit back and feel good about using a superior platform but that’s just existential posturing if it doesn’t translate into Marketing to a wide audience. For decades now, Pick people have felt comfortable and proud of being “The computer industry’s best kept secret”. That complimentary statement (magazine quote) was the kiss of death. The problem is that people ride that little wave of secrecy all the way into total obscurity.

These days the difference between “hot new” companies and products, and dead ideas and failed businesses, is less about technology and more about evangelizing, the social graph, and viral marketing. You don’t need the best product to be successful. You do need to be very good at telling people you have the best product. Actually being the best just helps to keep you in that position – but you need to get there first.

The way to get MV to new audiences is to put heavy focus on extensibility, and accessibility from common mainstream languages and frameworks. When MV is as easy to use as MySQL from PHP, Java, Ruby, and .NET, we can get people to use those tools for their client code while starting to appreciate the value of BASIC on the back-end. “It _is_ easy” you say? Sure it is, but if only Pick BASIC programmers know how easy it is to use MV with other languages, then what good is it? That message needs to get out there among Java and PHP and .NET developers who are looking for a better database than Oracle, SQL Server, or MySQL.

All of the MV platforms, like Universe, Unidata, Reality, D3, jBase, and Caché have the potential for being the “hot new” data platform, especially when NoSQL is now in vogue. The problem (with some/most of these platforms) is that they’re all too sprawled out into the OS, with files and registry entries littering the end-user systems. QM and OpenInsight have the smallest footprint of the MV platforms, and in my opinion they are best suited for introductions to the mainstream. Before vendors of other platforms get all uptight about this, once people are using these platforms, companies will seek alternatives. There are many examples in the tech world of companies getting familiar with one platform, and using that as the starting point for further inquiry. This opens the door to all of the other MV platforms. The mere presence of competition serves to lend credibility to any platform. The point here is that in the name of survival the MV vendors should support strong marketing efforts for QM and OpenInsight, making sure that their own platforms figure prominently for that second wave of inquiry.

I don’t want to spend too much time talking about the faults of the platform but we all know that installation is usually a pain, command-line maintenance is a pain, and there are way too many esoteric nuances for the average application developer to manage. But the MV platform is really no more or less complex than its relational competitors. The difference is that the relational guys expose a simple to use interface for app developers and they leave database administration details to “DBA” specialists. In the MV market we pride ourselves on not needing a DBA, and then we expose users to all of the complexities that can scare people away.

Regardless of the pros and cons of the platform or of individual MV DBMS products, the bottom line is that we simply don’t compete with the relational market. Even the MV DBMS companies themselves try to sell their products to the MV community based on “features”. It’s not about “features”. Features come and go. What we need are uplevel business partners who will keep the platform alive into the next generation. We need ways to educate new developers and to improve the job market for people who have invested in their MV DBMS skillset.

We’ve seen what happens when MV DBMS companies do no marketing outside of their own community. The market shrinks.

We’ve seen what happens when the goal of each MV DBMS provider is simply to steal business from other MV DBMS providers. The market shrinks.

These companies need to collaborate to introduce the platform to new audiences. Hey, c’mon, what are you guys waiting for when the rest of the world is on a “we hate SQL” kick? Your job doesn’t get much easier.

The best example of collaboration I’ve seen in many years was at the TigerLogic conference in November of 2009 where Rick Koe of TigerLogic and Susie Siegesmund of Rocket Software stood together to talk about collaboration to market the Pick/MultiValue platform to new audiences. It sounded like they finally “get it”. It sounded like two high-profile companies were going to take a lead role and bring together other MV DBMS companies in the name of common interests. It sounded like the Spectrum Manufacturer’s Association. Mike Ruane of Revelation Software was in the audience too. Wow, what a power trio to get the ball rolling.

Well it all sounded good but it never went anywhere. Until we see it happening, we’ve all just wasted a lot of time. The resources are available to do some very good things for the platform and for ourselves. What’s lacking is interest, determination, and maybe belief that it’s possible. Our own lack of conviction is what continues to erode this market, from innocent statements about lack of ability to do a nice front-end, to lack of marketing on the part of companies that you’d think would be motivated to tell the world about the gem they have to sell. No one will put their faith in a platform where the people who use it and sell it don’t seem to really have faith themselves.

Personally I believe this malady can be cured with some focused effort. For many years I’ve had much of my enthusiasm beaten out of me, but not my fundamental belief that the platform is truly capable of competing in the mainstream market. I just wish the people with the resources to do this had the same enthusiasm, belief, and fiscal interest to really take this software where it deserves to go. At the end of the day, we are in business to make a profit, no?