Universal Data Option. Was: Oracle NVL equivalent ?

clem@informix.com (Clem Akins) wrote: :>> Peter inquires: :>> >We currently use NVL to return a value (0) in place of a null in a :>> >select statement. Is there any equivalent in Informix? :>Clem wrote: :>> There is an exact Oracle equivalent scheduled in version 7.3, available :>> in the Spring of '98, and no supported equivalent prior to that. :Avi asks: :>And what about INFORMIX Universal Server? Will that also include an :>equivalent to the NVL function? : :This feature is also on the slate for inclusion in the Universal Data Option :of the Dynamic Server (currently unpronounceably called IDS/UDO) but not :at the same time as the 7.3 release. Current plans call for much later :in the year (though that is a long way off and subject to a lot of :change.) : :This question opens a whole range of issues about the demand for the :Universal Data features and the proper focus of Informix Corp on the :version 7 versus version 9 products. When the Universal Server was :initially announced the company wholeheartedly and unashamedly devoted :all resources to its completion. Analysts and (more importantly) :customers did not view this favorably, called Informix "too visionary", :and generally did not feel that the additional capability was worth :nearly double the price. Following the unfortunate coincidence of :timing with the accounting and business practice debacle Informix :implemented several profound changes, among them a shift in focus :from "visionary" products to the flagship version 7 line. This shift :comes at the expense of the migration of new features to the Universal :product set. : :The good news is that these changes are subject to customer input. You :can influence this direction by buying the Universal Data Option, and :by asking the sales force for information on when new features will be :available and indicating your willingness to buy when they are. This :kind of feedback has great meaning to a company struggling to build back :flagging sales. : :Contact your sales representative and Informix via the web site to :register your desire for the Universal Data Option. I have. (First Clem: "You" below isn't you, but Informix. I realize your level of understanding, but you will have to excuse me for letting you represent Informix this time.) However Informix show very bad judgement in this case. You create a significant chicken & egg problem for yourself. The UDO doesn't have the basic features we all need and want. The UDO does however have the advanced OO features we also all need and want. A problem is however that many still don't know they need or want the UDO features. Now Informix want user input on this need. They even want indication of willingness to buy. To buy something relatively few users know how usefull it would be to them. How much input do you expect? Now you could say that Informix is becoming more a market driven company. They are looking even more at what users want and supplying that. The big problem is however that generaly can't demand something they hardly know what is. Even supposedly advanced users like those who buy databases can't be expected to do this. If you want you may look at c.d.i. There are a few top people who know what they want and need. Large numbers don't however. This is a too well known fact of marketing that it should have escaped Informix to such an extent as it has. The UDO question adressed here is not the only case showing this. Another effort still under way to make Informix do something new also suffers significantly under this misunderstanding. What Informix should have done is analyze their user base, including future new users, to find out what they realy need. This should of course include user input, but must go much further. It's up to Informix to know what we "need" in the future. You should tell us, not the other way around. "O", you say, "but that was the error committed with the UDO option in the first place. Informix did just that and we where wrong. Now we are on the right track again." That's not it. The UDO option was the right thing. You did however create several unnecessary problems for yourself. First, and by far most important, you tried to charge an innordinate price for it. That might have been fine if it was a product for a few select customers and you could get them to pay enough to support a special engine. That's however not the bussiness you are in. You are in the general database business, and have to stay there to be successful. That means you need basically one engine, possibly with some options. The key point is that you are in a volume market. In that sense you are now on the right track. The UDO "option" is however not a true option as you have shown. You can't add it to a running installation. It's still a special version of the engine. You have to install it as a separate thing. If you want to use any of it's features in a database it have to be created within this engines separate dataspace. True you can use this for "regular" databases as well, but as you say then you forgo the 7.3 features that are also important. To fix this, if you ever will, is at least a year away, perhaps more as I now see it with your current plans. Even if it's only half a year away you have a significant problem. I, and I would expect most others, can't accept the problems of two engines just to utilize the UDO features in some cases. If we have to wait a long time you will have lost a significant market advantage. What should Informix do? Obvious: 1. Charge the same for IDS/UDO as for the regular IDS engine. We would probably accept the lack of 7.3 features for some time. At least we would have had that option. 2. Create a Workgroup and Developer version of the IDS/UDO engine, again with the same prices as the current regular IDS engine. We could then use the UDO for all installations. 3. Port all the 7.3 features into the IDS/UDO engine. 4. If 3 can be done well enough to attain the same performance as the 7.3 IDS engine will have, you may then drop the regular IDS engine from your product line. I hate to say it, but there is an option for charging extra for the UDO features: After point 4 above is done you could again create a special version without the UDO features. This time however it could be upgraded to UDO by a simple install of additional software. After this install all data would be intact, no changes need be done to anything allready running, and you could start using UDO features on any of your databases. I dislike the very idea of an extra price for the UDO features however. We have now come so far that the UDO features should be part of any serious contenders offerings in the RDBMS field. As this is not yet the case, and it will take some of the others a long time to get there, Informix would have a significant marketing advantage. You are currently squandering this simply by charging extra for it. It's also very simple to show that you can't possibly earn any significant ammount of money by charging extra for the UDO features. The small percentage of customers who elect to buy it now can't possibly make up for the