HTML vs Windows clients

[iansdigby]

Apologies if this topic has beeen discussed here.

I am wondering whether my TM1 skills will still be marketable in 5 years' time, given IBM's HTML/Java-based strategy for TM1 clients, without entirely embracing the new way of doing things and learning in depth about a much larger stack of TM1 clients. I'm getting on and have about 6 years of working life left.

Am I deluded in drawing a comparison betweeen the current IBM strategy and that of MS with VBA vs .NET? Microsoft, for good technical reasons, wanted Office development to be done with .NET, Visual Studio, VSTO etc. It effectively stopped development of VBA around version VB6 and yet today there are thousands if not millions of business-critical VBA models still in place and MS still bundles it in every new version of Office.

Will we ever see complete redundancy of the Windows client - based models in TM1, or is the critical mass of installed models too great for the wholesale change to HTML/Java? Or, has IBM succeeeded in the strategy and I'd better go back to school right now?

Does anyone have an opinion?

[qml]

Making predictions is very hard. Especially about the future. <ba-dum-tss> Having said that, I don't see IBM moving away from Excel integration. Sure, there will be more emphasis on web-based interfaces, but at least some of the old-school skills will still be transferable. CAFE is clearly inheriting more and more functionality from Perspectives.

On the other hand, learning new tricks will also be required for professionals who want to stay relevant, just like it has been with every major version of TM1 I have ever worked with. But it has been very gradual, hasn't it?

People usually overestimate short-term changes and underestimate long-term changes. I don't think IBM has the will or capacity to shake things up massively in a blink-and-you'll-miss-it kind of way in the next 5 years, just like it didn't manage to do it in the last 5 years. So you're probably all right if you just want to coast for a bit longer.

[David Usherwood]

@Ian, I'm somewhat nearer than you to the milestone you mention

You seem to be focusing on TM1 frontending and to be concerned that IBM will be dropping the Windows client - which must be (I believe) Perspectives. And yes, IBM have laid out their plans to 'sunset' Perspectives, but the plans are to replace Perspectives by CAFÉ. The latter started out life as a BI to Excel tool but now supports TM1. It has a way to go to be a complete Perspectives replacement but DBRW and (some) other formulae are in the current release and Active Forms and Action Buttons are expected soon. IBM know that Excel isn't leaving the finance community any day now.

Apart from Excel, the frontends will be web-based (shouldn't really be a surprise). The reports from Vision describe something called 'Planning Workspace' which is the PRISM project rebranded. The first iteration will be consumption (dashboards, of course ) and budgeting - but all the other tools and tasks will (eventually) migrate, specifically development (bye, bye Performance Modeler - no tears shed there I suspect) and Ops Console (ditto). CAFÉ (maybe obviously) does not include Perspectives' developer functions. I don't know what kind of customisation will be offered in all of these and what language/tools will be offered - but you should look at the new APIs in 10.2.2, launched in May 2014 - REST/ODATA for clients and ExecuteJavaN/S for TI. So yes, Java for servers and Web Services for clients is a pretty safe bet.

All that (OK, apart from ExecuteJavaX) is about the client side. The server side (rules and TI) is still pretty d@mn good, and (I think) skills related to those will remain in demand. For Excel, much depends on what IBM do to port the 'Excel API' to CAFÉ.
Don't get your pipe and slippers out just yet.

[iansdigby]

learning new tricks will also be required for professionals who want to stay relevant

Indeed qml, and I'm up for that. It really helps to hear your thoughts and experience. Coasting sounds attractive, but I suppose we should not 'go gently' and all that. Thanks.

Don't get your pipe and slippers out just yet.

I won't David, and thank you very much, not not only for the encouraging words but also for a revealing synopsis of IBM's road map. A lot of that I had no idea of.

bye, bye Performance Modeler - no tears shed there I suspect

...not a drop.

[Alan Kirk]

I concur with what both David and QML said to a large extent, except for the bit about how far some of you are from your sunset years. It's taken a lot of time to get this forum to where it is and we can't afford to have any of the stalwarts aging and retiring.

So stop aging, all of you. Stop it right now.

David made a valid point regarding a lot of TM1 skills being about server-side model building, including rules and TI. There are changes coming on that front as well apparently - talk of feederless rules and such coming out of Insight 2015 - which will also change the way models are built. But as QML was indicating, this isn't going to hit with the intensity of a tropical rainstorm. 6 years from now there will still be TM1 customers who need models built.

The client side is arguably a good news story since some within IBM has recognised that we were being buried in them and are starting to streamline them into two; Prism-as-was and CAFE. CAFE is pretty easy to get your head around. If you know Perspectives you can be working constructively with CAFE within a day. Or I was, and I doubt I'm exceptional in this respect. It will get slightly more complex as it inherits functionality from Perspectives, but it's far from uncharted waters for someone with your level of experience.

You'd probably need to learn whatsisface Workspace - ah h3ll, I'll just keep calling it Prism for now - but I'm not sure how arduous that would be. I suspect not exceptionally if you have a session set up that you can use for "playtime" and experiment with.

The good news is that you won't have to worry any longer about learning Muddler or Insight, since they're evolutionary dead ends.

The biggest learning curve you would face is the RESTful API one, which is a very different experience to the classic ones that we've masochistically applied ourselves to over the years. But I don't think that this is mandatory; many developers do just fine without ever touching an API except maybe the Web one occasionally.

I look at it this way; many of us got into this game because we like learning cool new technologies anyway. It's often as much recreation as work, although I'd include the caveat in my case that it's only fun if I can get something useful at the end of it. (Which is why I'd rather, say, write C# code than fuddle around with Linux.) So given that you won't need to learn everything about every new thing that comes along, don't look at it as 6 years of coasting, look at it as 6 years of potential fun.

Or more than 6 years, since I remind you that further aging is strictly forbidden.

I won't David, and thank you very much, not not only for the encouraging words but also for a revealing synopsis of IBM's road map. A lot of that I had no idea of.

It's not a bad idea to skim these two threads since they provide some further discussion about the roadmap as we understand it:
Insight 2015
Future of TM1 UIs and APIs (Was TM1 C++ API issue)

[iansdigby]

many of us got into this game because we like learning cool new technologies anyway.

Yes!

Thank you Alan for your thoughts. I feel loved and re-invigorated by all this encouragement and will not now REST until I'm 'API.