TM1 Forum

Guys,

As we all know IBM have at times been on this forum and have been of great assistance. May be it's time we gave them something back.

If you could ask the development team a question what would it be??? If you could raise a suggestion what would it be???

Please keep in mind that IBM will more than likely read this thread take your point seriously so please, keep your posts to the point and sensible.

Happy hunting guys......

Why does Replication in 9.5.2 not work? Do many clients use it?

Why do you need to click Ctrl C lots of times to copy things in TM1?

Below are a couple of suggestion im sure there are lots more
For Drill through once it has drilled to the Relational Drill Through screen there currently is a copy and select all. Can there be a export to also.
In TI it would be useful to be able to display a Message Box.
When you click on Chores in the display properties Window it would be useful to see the time the Chore is scheduled for.

CC7 wrote:Why do you need to click Ctrl C lots of times to copy things in TM1?

An IBMer explained this one to a bunch of us over beer the other week (see Kyro's sig for more details on the beer). You only have to hit it once, but you have to make 100% sure that you release the C before you release the Ctrl. It has to be Ctrl key down, C key down, C key up, Ctrl key up.

I have been doing this since then, and it does work on only one press.

What are the plans (if any) on enhancing the TI GUI and replication?

On replication specifically, ideas on improving maintenance once replication is setup. The ability to remove replication to certain cubes rather than having to start again or fiddling with data in control cubes. Ability to replicate just part of the replication links rather than all at once. When choosing a cube to replicate and including its dimensions, ability to replicate the dimension attributes also (without seperate replications for the attribute cubes).

What are the plans (if any) on enhancing the TI GUI

Seconded.

Autocomplete would ease my early-onset-carpal-tunnel-syndrome, better known by it's medical name DimensionElementComponentAdd.

-> At what point will they consider Cognos fully integrated with TM1, and how much further down the road is that?
-> When will we see improvements to TM1 Top, and user/system management across multiple instances of TM1. Vendors are innovating here, where's IBM's contribution?
-> TM1 Top - often everything locks up and we can't tell exactly why. Visibility into who is locking what and which users are waiting on which locks?

Christopher Kernahan wrote: -> When will we see improvements to TM1 Top, and user/system management across multiple instances of TM1. Vendors are innovating here, where's IBM's contribution?
-> TM1 Top - often everything locks up and we can't tell exactly why. Visibility into who is locking what and which users are waiting on which locks?

I definitely agree with you that TM1Top has to be improved. As the state and function refresh so fast, sometimes I have difficulty to identify which cube this particular PID is related to. It will be good if there is a separate column to tell that this pid is related to xx cube and yy view. There is situation when a user open multiple view within a cube and the view hang and I do not know which PID should i be cancelling.

Anyway is the cancel function on TM1Top working ? The PID is still there after i try cancelling it. I am using CE 9.5.

Here's mine. Third party utilities like Vizier and TM1 Tools have identified substantial gaps in the basic functionality. When are (or, indeed, are) these gaps going to be addressed in the main TM1 product? Specifically:
- The already mentioned multiple times TI interface. Most of the limitations can be worked around in one way or another; for example if you want search and replace, albeit one tab of code at a time, you can do it by moving the code out to Notepad. Or, you could use WordPad if you want to be able to actually see the text and punctuation (such as the comma that looks like a period, but isn't or vice versa) on a high resolution screen because the code, in its unchangeable and small font, is reduced to microdots. This is a trivial bell or whistle, because working around these problems may take only a couple of minutes here, 5 minutes there? Now multiply that by the number of developers. And multiply that by the number of projects that they do. How many tens, hundreds, thousands, even, of hours have been wasted over the years because of TI's anachronistic interface?
- Hot promotion. Similarly how many hours (tens, hundreds) have been wasted by having to shut down a production server, copy the files of a new cube / process / chore from a development server, restart the production server?
- The functionality provided in Bulk Paste? Zap Formulas? Trace formula? These are things that every TM1 user needs to do periodically and yet they remain the province of third party apps.

As another arm of this, what's the state of development of the various APIs? Will gaps in the functionality of the "classic" API ever be plugged, or is the development team merely adding on functionality to support new features? (Or, indeed, are they even going to continue to do that since VB6 and plain old C are rapidly becoming "yesterday's languages" and the only thing that's keeping the classic API on usability life support is VBA?) Will there ever be further development of the .Net API to make it useful for something beyond pulling a number from a cube, given that .Net underpins a lot of web applications? Will there be ongoing support and development of the Java API?

There are any number of questions that could be asked, but my greatest area of interest is in the basic functionality which, aside from the calculation engine itself, is the foundation of the whole product.

Or, you could use WordPad if you want to be able to actually see the text and punctuation (such as the comma that looks like a period, but isn't or vice versa) on a high resolution screen because the code, in its unchangeable and small font, is reduced to microdots.

Agree - I thought my eyesight was bad in high school. Now I'm using binoculars.

Note that TM1 team members are not able to comment on future product releases.....

I have directed various TM1 folks to this thread so you will be heard.

dr.nybble wrote:Note that TM1 team members are not able to comment on future product releases.....

I have directed various TM1 folks to this thread so you will be heard.

They aren't... but the intention of this thread is to gather some information that will be used in making some representations to IBM. Of course, one would hope that they're already keeping an eye on the Enhancements forum , but the intent of this thread is a little more specific.

Simple.

Perspectives Wide Area / VPN Performance, please make it work better.


(ADMIN : Small mod made to make this polite! )

(Not sure the mod quite conveys the exasperation of why 30ms latency brings an enterprise app to its knees as well as the original three letter abbreviation, but I'll abide!)

wysiwyg TM1 web?

and added support for text boxes, shapes, 2007/10 cell colour blending.

but most of all more flexability in the display of action buttons. Why cant I attach a TI process to a cube shape? or maybe attach it to an arrow linking two cubes? Its just a bit more intuitive than a button or hyperlink.

I know its not everyones cup of tea but TM1 Web does allow you to have some white space and properly explain what a contributor needs to do. Especially for users who infrequently use TM1 and you dont want to much around with their excel environment.

oh and you could do some really smick demos with a little more functionality

Stex

Well, I'll ask for the same thing I asked for in the last Xmas Wishlist (of around 4 years ago) - can we have a unique object to represent a view (such as a .que object) that can be used as a TI data source rather than a vanilla user viewable view?

Same goes for subsets used as TI data sources to a lesser degree but all the ZZZ_ views or dynamic creation and deletion of views has been and still is an irritation of the highest order.

Pretty sure I raised an enhancement request around seven or eight years ago about this

garry cook wrote:Well, I'll ask for the same thing I asked for in the last Xmas Wishlist (of around 4 years ago)

Try 6. The legendary Christmas Wishlist was kicked off by the BigDster at 10:19pm (my time), Wednesday 7 December, 2005. A day that will live in infamy, only because of the number of requests on it that were never implemented.

garry cook wrote: - can we have a unique object to represent a view (such as a .que object) that can be used as a TI data source rather than a vanilla user viewable view?

Same goes for subsets used as TI data sources to a lesser degree but all the ZZZ_ views or dynamic creation and deletion of views has been and still is an irritation of the highest order.

Pretty sure I raised an enhancement request around seven or eight years ago about this

You certainly raised it in that list on Friday 27 January, 2006. (Is anyone feeling old yet?) I'd second it.

(I probably shouldn't mention that your very first request on the list (8 December 2005) was for "a pair of blonde female Swedish masseurs". Ooops, just did. But if Applix wouldn't supply them, IBM certainly won't.)

I'd also like to add something that we've forgotten about, though the need for it has never really gone away; the second post in that list was from Steve Vincent, asking Santa for a step-through debugger for TI. Long, long overdue.

Hi all,

I have a suggestion about Tm1 views.
I would like to be able to get more control in views browser regarding the combination of dimensions on columns for example.

If I pile up 2 dimensions in columns in a view one being year, the other one being version.
The view will show the combination of all the selected elements in both dimensions ( minus the empty areas if I have zero suppression activated.) .

Version : Actual, Budget
Years : 2011, 2012

As a result in my view I will get :
Actual 2011 , Actual 2012 , budget 2011, Budget 2012.

Assuming I am in 2012,
and I want to check my actual 2012 versus budget 2012 and want to have only actual 2011 as a historical reference. ( Budget 2011 being now irrelevant in analysis).

Today, I would have to slice this view which is larger then necessary and customise the slice in Excel to get rid in this case of the Budget 2011 column.
Of course, I could build smart MDX excel reports to get there, or even simpler excel reports with dynamic column logics...or you might question the design of the cube in certain cases...

But I feel that allowing standard Tm1 users to perform these types of customisation directly in a view could bring a lot of flexibility and why not sometime kill some Excel outputs which are just formatted/restricted views...
I think that would make Tm1 views more flexible then Excel pivot tables on this particular set up.

Maybe a simple additional configuration menu in views on the model of the view filter option showing the column stack combination currently active and enabling the user to tick the combination he wants to keep in the display.

Column filter :
X Actual 2011
X Actual 2012
o Budget 2011
X Budget 2012


Hope this makes sense.

Kind Regards,

The focus on end user interface to the detriment of development tools is irksome (and not just as a developer, I've been in pre-sales and I understand the whole point about features that benefit 99% versus 1% and that ultimately the CFO wields the checkbook and they might be more prone to whizzy end user features). BUT and it's a big but, at the big organization end of town where you have big IT you also have a need for tools to be "enterprise ready" and that means more than a nice TI editor. It means real solutions to things expected of enterprise systems
- object promotion and synchronization tools for D-Q-P
- replication that really works and is reliable (including replicating dimensions without dummy cubes)
- out of the box or "snap on" support for code base (source) versioning
- support for non IIS based web server
- out of the box backup, redundant data mirroring, etc
- support for changing object names (cubes, dimensions, etc)
- client performance that is not crippled by WAN latency of more than a few dozen milliseconds

Add to that yes I would really like to see a fully featured, fully functional "development studio" environment as part of the base product. Yes there's Vizier and it's great but why isn't this part of the core product? Same goes for server monitoring, same goes for user and security management. By that I don't mean just fix those security dialog that take for ever to open (I mean really how about some caching and scrolling? No reason performance in these dialogs should be any worse than the cube viewer), fixing these would be nice but something like what FirstQuarter have developed would be much nicer still and should really be baked in.

The cube viewer. Let's face it, it is a bit early 90s in terms of UI. Having configurable or customizable default colors or view specific colors for titles, columns, rows, grid (Hello EV?). And how about a charting engine? ... and what about being able to build asymetric views or pull views together from different cubes (or parts of the same cube) in a "canvas" and have these objects communicate via "channels" so that when one title is changed the is reflected in all the linked views - same goes for a rowset or column set. Again not exactly a new concept.

Contributor. It's great (but somewhat limited). Adoption is limited to companies migrating from EP and likewise for consultancies migrating their business from EP to TM1. But it isn't exactly taking the world by storm in the traditional heartland of long term TM1 partners and "Excel centric" (mainly finance based) user communities. Why? Because (IMO) basic functionality and flexibility that is there in Excel and TM1 Web isn't there and this severely limits the ability to build an "application" (as opposed to a "workflow" or visual tree representation of task status actually) that can be intuitively navigated by end users. The main things that are missing:
- the ability to construct asymmetric views and put in "white space" spacer rows and columns
- the ability to combine views within a tab
- the ability to easily customize colors and fonts
- the ability to have additional selection constraints around view layout and title element selection (beyond element security). (See Enterprise Services for examples of how to do it)
- the ability to have event driven execution (i.e. running TI processes)
This means developing an application in Contributor is awkward. One very quick fix would be for Contributor to support TM1 websheets as tabs of an application. This would cover off all of the functionality gaps identified above, and so would make the interface much more usable and useful. But it would be a quick fix, I would like to see something more fundamental.

Hope that doesn't sound like too much of a rant. Merry Christmas!

All good points so far and some very well constructed posts. Keep them coming guys,

Jim.

I would like to see

1. The end of the symmetric dimension philosophy and the introduction of specialised time and measure dimensions that would allow a degree of consistency across applications and tighter integration with such tools as BI.

2. The introduction of Layered dimensions with constraints built in to the dimension as to how the layers interact. The ability to restrict a view of a dimension to a particular starting subset which generally would be layers.

3. The ability to post at C level may be too dramatic and allow monstrous complexity if not used corrrectly but taking the restricted view of a dimension should allow you to remake what is considered an N level for that dimension. Otherwise a facility that allows you to replicate portions of the dimension structure for N level posting would be good.

1 to 3 can all be programmed in TM1 but crop up again and again as problems which other OLAP tools have solved.

lotsaram wrote: It means real solutions to things expected of enterprise systems
- object promotion and synchronization tools for D-Q-P

This is the single simplest fix to implement - a feature to unload and reload single TIs from disk with a feature to repeat the initial load from disk of all .pro files.

It could be done in a minor release and would give IBM the biggest effect for the least effort.

I don't really have a question, but since this thread is heading down a wish list (aka thinly veiled gripes) path.... Here are ones that I consider major:

• Replication: It's really terrible. It needs to be fixed. No one should consider using it unless there are absolutely no alternatives, and even then I'd still think long and hard. Get a massive box and do RemoteApp instead. I don't even know where to start on the replication-specific fix list but my two big ones are that it needs to be much faster and it needs to be atomic so multiple reps can occur simultaneously.
• WAN Performance: Also terrible. Fix the chatty messaging protocol, make it out-of-order and asynchronous, and compress everything aggressively. We had to give RemoteApp access to all of our users since TM1 isn't usable over VPN, even in the same metro area.
• TM1 Web: Useful for only the very simplest purposes, and it's incredibly kludgy/ugly. I have to plan an extra hour after I build a template to get all of the columns sized correctly since what you see in Excel has nothing to do what how it'll look in the broswer.
• Installer: Why is it 500MB? There's no reason server and client (and web and contributor...) should be bundled together. I have to provide it to users all over the world and it can take hours to copy a 500MB file across some of our longer/slower links thanks to Microsoft's crappy Windows copy performance.
• Documentation: Errors, missing functions.
• TI Scripting: I know there is exactly zero chance of this happening, but it's ludicrous that TM1 has its own scripting language. It's embarrassingly limited. Move to something else, preferably C++ (or C#), and provide a library.
• Logging: There are a LOT of "ERRORS" in the logs that should be "INFO." e.g. end-user MDX failures. This might seem like a nit, but our SOX people make me track every single ERROR (as defined by the s/w) so this is a major PITA for me. Bonus points for actually documenting tm1s-log.properties so we don't have to go to support to make sure we set the right flags.
• Dev Tools: Need better ones. Debugger would be awesome, for example. Breakpoints and a stepper, that's all I need. It would also be nice if there were a way to track down performance issues that didn't involve trial and error.

BTW, for the record, I love TM1. I'm its biggest internal cheerleader. People LOVE trashing their Discoverer templates and using TM1 instead. But it sure is frustrating sometimes.

Matt