Hi Baker -- good points and questions. On 4/16/07, Baker Hughes <[EMAIL PROTECTED]> wrote:
If you don't want to use their own 'Zen' object stuff, can you use whatever you want? C#, VB, Java ?
Since Cache' has a development platform, like all MV systems have, there is a p-machine side (like all MV systems except jBASE have) in which their database code and all app code within Pick or MUMPS or their OOP that incorporates these runs. As with (must/all) MV solutions, this is not the same p-machine that runs Java bytecode (although historically coming from the same place). But they have a seemingly very well architected Java solution called Jalepeno, that I think, maybe, perhaps (I haven't read much and what I read was a few months back) permits one to access Cache' Objects specified as Java objects. So, you work with stored cache' objects (which are records in files to us U2 folks) like you work with any other objects, using standard java libraries for collections and such. You may also execute SQL statements against them, but you don't have to. This feature was so impressive, the story goes (as told by Intersystems folks, so IBM can jump in here to correct this if I have it wrong), that IBM chose Cache' using Java and Jalepeno over U2 or any other IBM databases for a call center project at IBM.
How does it like ODBC, ADO or other methods of getting to the Cache data?
THE method for doing SQL to Cache' is ODBC/JDBC and they have made it wicked fast from what I heard from customers in sessions at their conference. I am very impressed with their technical savvy as a company and their focus on performance and scalability, so I do not doubt that it really is incredibly fast to use ODBC. I haven't used ADO myself and couldn't spell it if you asked me to. If someone mentioned it (and they might have), it didn't stick in my brain.
We've all had our days hemmed into a proprietary environment.
Yup, now you are getting into the area where I wrestled myself to the ground (and won, I think).
I hope I'm wrong for the sake of my comrades like Dawn, but I haven't heard anything to make me think otherwise, but that your development efforts will be dependent on that 1 company, and tied to their survival.
I have experience with multiple database environments and know there is no such thing as database-agnostic applications through and through. The "if" statements in an application that attempts to work with more than one database are significant. Oddly, these might even be fewer in the Pick world (crossing platforms such as UniData and UniVerse, for example, even before they were jointly owned) than between most SQL-DBMS's, even though there are industry standards for SQL. So, I had to decide whether to do what many companies do and tie myself to a single database, making it a) hard to migrate to another if we had to and b) hard to have a good negotiation position with the database vendor should I need to. That is the reason that I am planning to use MV BASIC code, rather than MUMPS code, for functions within the software, first of all. Intersystems has almost a monopoly on MUMPS, but they definitely have competition in Pick. That doesn't get us all the way there, however, since there are the Classes, particularly those that correspond to the UI. When moving from one PIck solution to another today, the client-server or GUI is the most likely thing to not move. I will be tied to Intersystems for the UI, I realize. It is one of the costs when doing my cost/benefits analysis. I did not see that choosing UniData or UniVerse was going to do better in this area and I have what I think are good reasons for not using Oracle, SQL Server, or MySQL for the back end too (even if getting them on paper is difficult). So, I am addressing that in the relationship with the database vendor (ask me off-list if you are interested, but I really have no brilliant answer to this concern and it is one of my concerns). The good news for me is that the model for software delivery has changed. I will not be selling a product to multiple orgs who have to install in on their systems, each with different requirements or desires about what the DBMS would be. I can put whatever back-end on this that I think is good and provide 'web services' (good data exchange, at least) with related systems. The user can care about the business solution in an SaaS solution, rather than the technical components. So, I don't have to sell Pick or Cache', but do have to support the solution long-term and, yes, would be tied to a vendor at least for the GUI (AJAX) side of things, but can do the business logic in MV code, which has some portability, even if a migration would be painful (as if there were migrations that were not, eh?)
Just some questions: If it's so good then what vertical apps are porting to it? (from MV or RDMS spaces)
I had the feeling that those of us at this DEVCON are in the first wave of post-general delivery prospective customers considering migrating (after the initial "beta" folks). It is likely that only those of us who are not migrating (consultants like Tony and me and companies writing new software, like my new venture) would want me to give away their identity (they can speak up if they desire) since they have competitors and are considering this move to be strategic (perhaps also not wanting to alert IBM to the possibility if they are considering Cache'). I will tell you that Cache' is huge in the health care vertical, which is not my current area, so I was a bit concerned about that. My concerns were pretty much addressed by those in other verticals who seem exceedingly happy with InterSystems and whose businesses are thriving. Financials is another big vertical they are in, but they are very clear that they are a tools, not apps, provider and they will support application partners in any vertical with enthusiasm (and they really do have such enthusiasm, both for their product and for my business plan and interests).
Why are they going after the MV market so strongly ?
They bought out most of MUMPS and seem to have enough former PIck folks there who were saying that it was a similar situation of needing to take Pick apps into a broader environment with today's expectations for 24/7 but with our (PICK folks') expectations of big bang for the buck too. In other words, we think like they do and there are, I think, former pickies in their midst who said it was a good match. It also sounds like they considered attempting to buy out the Pick market like they did MUMPS a while back, but it didn't happen for whatever reasons. So, they are providing what they think is an excellent technology suite and it now includes MultiValue. They are clear with their story (seemingly changing it from some time back) that there is not Cache' for MV, but that Cache' now IS MV (as well as being MUMPS and OOP).
After you've written your killer app, and the buyer wants a different db, what are your options? (can you do a [somewhat] simple data/dict conversion and take it to some *other* MV database?
As you have seen from above, I'm doing SaaS, so I do not have that problem. They based their implementation of MV on UniVerse, I think, so that would be the easiest port in the other direction (from Cache' to UV), I would guess.
Have you seen any dollar figures on their R&D investment, compared to other 2NF db providers?
No, but I have "felt" their environment, talked to their VP of development and of marketing (among others) and I like their philosophies and strategies. Feel free to ask any additional questions about that off-list, but I honestly did buy into their corporate big picture and their operational means to achieve their strategies. As with any Pick "manufacturer" they are successful if their application partners (e.g. VARS) are successful. They make that clear with their creative approaches to pricing that aligns with your business model, for example. I hope that helps. --dawn
-Baker -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Dawn Wolthuis Sent: Monday, April 16, 2007 2:36 PM To: u2-users@listserver.u2ug.org Subject: Re: [U2] U2 / Cache I should perhaps put this on the U2C list, but I'll second some of Tony's impressions of Cache'. I am starting a new venture (adventure) and have selected Cache', using their MV tools as well as their AJAX tools, for the platform. In my case, I have done a few talks here and there in the past couple of years, with my primary focus in two areas: MV and AJAX. It was shocking and delightful to see that there was a company spending a large amount of its R&D dollars on two primary projects during this time: MultiValue implementation and AJAX integration. So, I had to take a look. Their MUMPS technology (with objects added, just as they added object features to MultiValue BASIC) is renamed Cache' Object Script. It is at least as unimpressive at first glance to a non-MUMPS developer as Pick might be to a non-Pick developer. But they set up the MV implementation so that, in theory (I'm guessing there are exceptions), anywhere that you can write MUMPS code, you can now write DataBASIC (Cache' MultiValue BASIC). Very cool. Cache' classes can be developed, looking similar to Java and other object languages, that generate MultiValue dictionaries. These classes then work with the rest of their AJAX implementation (named Zen). Any Cache' class with methods in it can have those methods written in MultiValue BASIC. As a sometimes Java developer, it is pretty interesting to see this blend of OOP and MV. Feel free to ask me any questions about the product or my decision on this, if desired, especially as we get into more hands-on with the product this summer (the devil is in the details and all). Of course, there will always be a place in my heart for U2, but I had to make the best decision for the new software we will be building and InterSystems with their Cache' offering that now includes MultiValue just "feels" more like the future, with their 24-7 approach, scalability etc. There was really good energy at the conference and in discussions with key Intersystems folks. I like the U2 group too, but they do not have the same level of corporate-wide backing for their MV product that Cache' has, so the innovations are not in the same place either. Cheers! --dawn -- Dawn M. Wolthuis Tincat Group, Inc. tincat-group.com Take and give some delight today ------- u2-users mailing list u2-users@listserver.u2ug.org To unsubscribe please visit http://listserver.u2ug.org/
-- Dawn M. Wolthuis Tincat Group, Inc. tincat-group.com Take and give some delight today ------- u2-users mailing list u2-users@listserver.u2ug.org To unsubscribe please visit http://listserver.u2ug.org/