TM1 V10.1 Release; New features And issues

[Kyro]

Odd, I can't seem to find "Applix.TM1.API.dll" in the bin directory in TM1 10.1 - has this been depreciated?

Edit: Even more strange - This does come with the installation but only exists in the TM1 Web Directory.

[Alan Kirk]

Odd, I can't seem to find "Applix.TM1.API.dll" in the bin directory in TM1 10.1 - has this been depreciated?

Edit: Even more strange - This does come with the installation but only exists in the TM1 Web Directory.

I don't think it actually matters (from a technical point of view) where it is, but I'd agree that (from a practical point of view) someone has screwed up and had it installed into the wrong place. The oh-so-useful .Net API .chm file appears to have vanished as well, with no "real documentation" replacement that I can see.

[jim wood]

Ouch. I know of projects that I've worked on that this will be a real problem for. Good spot,

Jim.

[Alan Kirk]

Ouch. I know of projects that I've worked on that this will be a real problem for. Good spot,

I think it would be a problem even had it been left in the original folder (at the main application level); the entire default install path has changed since it was oh so vitally necessary to include an "ibm" folder above the "Cognos" folder. And just to top it off, the install package doesn't update the Path environment variable to help applications find the location of the API libraries (including the classic API libraries).

[Alan Kirk]

This is interesting, and will cause a bit of a change in our understanding of dimension maintenance. Naturally this wasn't mentioned in the NFG, but hey, where's the fun in introducing new features if you actually tell the users about them?

There are a bunch of new dimension maintenance TI functions which have the word "direct" in their name, such as DimensionElementInsertDirect.

The default means of editing a dimension in TM1 is to use a whole-copy editing pattern. In that pattern, an editing copy of the dimension is created, edits are applied to the editing copy, then finally the actual dimension is rewritten using the editing copy as a template.

Oh wow... actual documentary confirmation of what we've been assuming. It only took what, 12, 15 years...

TurboIntegrator supports whole-copy editing automatically whenever dimension editing TurboIntegrator functions (like DimensionElementInsert) are used in the metadata tab of the process. TurboIntegrator automatically creates the editing copy and applies editing operations to it, then rewrites the actual dimension at the end of the Metadata procedure.

Direct edits are different in that no editing copy is involved. Instead, the operations are performed directly on the actual dimension.

These are intended for small updates to the dimension orrrrr... can be done in the Data tab, removing the need to have the TI loop through the data source twice (the Metadata tab can be left blank) and therefore substantially improving performance.

If it works as advertised this is quite a good feature.

For those who stumble upon it, anyway.

But a very good improvement nonetheless. (I'd like to leave it at that but based on the previous couple of postings I just know that there's something waiting around the corner that's gonna hack me off...)

[Steve Vincent]

Interesting. I wonder if that would also avoid the issue caused by aliases where if the structures are changed enough, the element number of the old element no longer matches the same elements' number in the rebuilt dimension? That usually causes a TI to fail as it thinks it has duplicate names, the only way to solve it is to seperately nuke the aliases 1st, rebuild the dim, then apply the aliases again. It also totally screws replication in the same way, of which there is no sensible workaround...

Of course there is a danger that the TI may go wonky part way through processing. Be interesting to see if it can maintain the data in a cube in such cases...

[rollo19]

I guess prototyping gets easier/faster and can now involve Joe Analyst but no replacement for skills in a larger production environment. e.g. the documentation suggests you should check all feeders (comment out and re-write if need) and re-order rules appropriately. Interesting new concepts/rule syntax with the auto-generated rules.

Agree those new smaller grey/blue icons don't help with usability. Perhaps the IBM UI people forget they were wearing these:

At last - the TI scripting editor gets a makeover..

[Alan Kirk]

I guess prototyping gets easier/faster and can now involve Joe Analyst but no replacement for skills in a larger production environment.

Amen.

Agree those new smaller grey/blue icons don't help with usability. Perhaps the IBM UI people forget they were wearing these:

Heh, I liked that one. I find the icons easier to identify now that I know what I'm looking at but they're still not as easy as the old N and S. Somewhere along the way IBM forgot that the whole purpose of an icon is to provide the user with instant recognition of what it is that they're looking at. It's still a design fail.

At last - the TI scripting editor gets a makeover..

In Performance Modeller it does; it doesn't in Perspectives. I had yet to look at Performance Modeller. Oh, right, now I know why!

Unlike Web, which I haven't been able to launch at all yet, Performance Modeller (or perhaps I should say Applications) did launch after I entered the butt-ugly URL address. "Intuitive" is not the word that came to mind and it certainly won't be the word that comes to mind for newbie users who will be confronted with (to them) unintelligible messages about things that are missing from their tm1s.cfg file. (In the sample databases, no less. The sample databases that Iboglix provided and which you would therefore expect to be configured to run with Performance Modeller.)

So what happens when I go to launch Performance Modeller? It asks me whether I want to install it. Yes, I have to freaking install it first. Gee, here I was thinking that I had selected all of the options during the initial installation and that it should sodding well already be installed. But no, I have to save the .msi somewhere and run it. At which point Norton warns me that fewer than 5 users have installed it and with big red flashing lights almost breaks out into a chorus of "Don't Pay The Ferryman, Don't Even Fix A Price, Don't Run This Installer, It May Not Get You To The Other Side!". I tell it to go ahead anyway. I'm now at the point where it's asking me to open provagent.cogrcp_modeler which it's suggesting that I open with IBM Cognos RCP Application Updater. Sitting... watching... hard disk spinning... nothing... still nothing... oh, splash screen!

OK, I'm in. Oh fabr'us day, it actually has search and replace, comment/uncomment and a few other niceties. Firing up this thing just to get to the updated editor should be lots of fun, rather like firing up a 44 class diesel locomotive to drive down to the local shops, but once you're in it'll certainly be an improvement over the Perspectives one.

Just to return to the installer for a moment, let's see what we've got, shall we? I mentioned some of these to Support earlier today:
- It can't uninstall a previous installation, you have to do that manually (and it doesn't even give you a warning that you have to do it manually, you have to read the lengthy installation guide to know that);
- It puts the sodding application into a different path because the install just has to be branded with the IBM name;
- It still puts data files (the sample databases) into the Program Files path, a practice that everyone should know by now (and certainly IBM should because I reported this isssue to them years ago) is not recommended;
- It makes NO effort to update the Path environment variable for those using the API. (The IBM service rep kindly pointed out that the fact that you have to do that is in the (classic) API manual, to which I unkindly pointed out that other applications seem to be able to do this automatically without requiring the users having to manually hack their Windows settings every time they do an upgrade... of course, most applications don’t change the sodding path with each version upgrade in the first place);
- It omitted the .Net API documentation, meagre as it is, and now;
- Despite me telling it to install everything I still had to manually run a sodding .msi just to get some of the functionality in place.

So what exactly did this vaunted new installation package bring to the table?

Oh, I know... a pretty new face with different icons.

(I'd like to leave it at that but based on the previous couple of postings I just know that there's something waiting around the corner that's gonna hack me off...)

You know how some people hate to say "I told you so"? Not me. I told you so.

Still, I'll be interested to try out the new editor and see how it performs.

[Alan Kirk]

Still, I'll be interested to try out the new editor and see how it performs.

OK, first off in Performance Modeller (which I refuse to spell with 1 "L" because where I come from that puts a red squiggle under it) the chore icons definitely don't change when a chore is active. Steve Vincent, this one's for you.

There's another cute quirk. You know how there are some chores that you don't want to schedule, ever, so you'd be quite happy to leave the schedule as 0/0/0/0 but you can't because the chore dialog insists that you enter a non-zero value? Well, that one's still there.

But you know how you would put in 999 days as your way of saying "Get out of my sodding face and get on with it"? OK, you can't do that in Performance Modeller. If you set (only) the days, you can only set a value from 1 to 365. (Of course even a P.O.S. like Quicken can do schedules which are annual (so leap years (oh! like this one!) don't affect the schedule), weekly, fortnightly, nth Thursday of each month, etc, but not TM1 chores. But then Quicken scheduled transactions are done in local time so what would they know?) Where was I? Oh yes, you can still set 999 days in ye olde Server Explorer but if you try to work with a chore like that in PM you get an error message telling you that the input is incorrect. (Not what the valid range is mind you; I discovered that from trial, error and guesswork though perhaps it's in the manual somewhere... just that it is invalid.)

Why the difference? Probably a "left hand/ right hand" thing. What was that thing we were saying about "same and consistent functionality across all interfaces"? Never mind...

[rollo19]

Yes the rich client install after an install when all on one laptop.. But this way any client can pull down Cognos Insight/DESKTOP or Performance Modeller via the Java app server (a web browser) without IT having to install something.

I would have thought the installer could do the un-install but no. After some messing around (cleaning out the registry) I discovered the un-installer in C:\Program Files\IBM\cognos\tm1_64\uninstall Assuming it works..

Still no joy with excel web sheets and TM1Web - posted on another topic.

[Andy Key]

OK, first off in Performance Modeller (which I refuse to spell with 1 "L" because where I come from that puts a red squiggle under it)

I though I was being helpful when I raised the single L as a bug during beta testing, especially as it is inconsistent and sometimes uses 1 and sometimes 2. I even checked that the spelling was incorrect for the US. As you can see, I didn't have much success. Unless they updated everything to use 1...

But this way any client can pull down Cognos Insight/DESKTOP or Performance Modeller via the Java app server (a web browser) without IT having to install something.

That's as maybe, but if you are using PM you will still need Architect installed locally to do all those tasks that PM doesn't do.

[Alan Kirk]

OK, first off in Performance Modeller (which I refuse to spell with 1 "L" because where I come from that puts a red squiggle under it)

I though I was being helpful when I raised the single L as a bug during beta testing, especially as it is inconsistent and sometimes uses 1 and sometimes 2. I even checked that the spelling was incorrect for the US. As you can see, I didn't have much success. Unless they updated everything to use 1...

Actually... according to Oxford it is correct for the US, just not for most of the rest of the world.

This is the rule for "World English":

Spelling rule
Double the l when adding endings which begin with a vowel to words which end in a vowel plus l (as in travel): (models, modelling, modelled).

And this is the one for US English:

Spelling rule
Do not double the final consonant when adding endings that begin with a vowel to a word that ends in a vowel plus a consonant, if the stress is not at the end of the word (as in target): (models, modeling, modeled).

So I wasn't saying that they were "wrong"... just that they are from our perspective and I'm not going to adjust my typing or irritate my spell checker by typing it that way. (Any more than the use of "color" in VBA is "wrong" or the letter "zee" is wrong, though I'm prepared to make the case for the annoying MMM-DD-YYYY format on the basis that you either drill from the specific (day) to the general (year) or the general to the specific but you don't hash it up like that. ) However if they have used both names then that's wrong but given the slack nomenclature standards in there it wouldn't surprise me.

Aside from which it's a meaningless name that'll probably be changed in the next version anyway. They like to do that.

(Half the time I have to stop myself from typing "Performance Manager"...)

It's that old chestnut again about "two nations divided by a common language", a pithy saying which was ascribed to Churchill (as most are), George Bernard Shaw (as most are) and Oscar Wilde (as most are) but surprisingly for once not Mark Twain. I'm pretty sure that George Patton said it once so I'll just credit it to him.

[Alan Kirk]

At last - the TI scripting editor gets a makeover..

It's a constant source of inspiration to me when someone does the bare minimum that needs to be done, and calls it "Progress". So consider me inspired.

10 years late, we have Find and Replace. Ah, and what an implementation it is. First, the usual MS Office (and many Windows applications) key combinations of [Ctrl]+[F] and [Ctrl]+[H] haven't been implemented. You can only get the dialog from clicking on the little magnifying glass icon which looks not quite like the magnifying glass icon that represents "Goto Line". But who cares if they're intuitive, as long as they're pretty. Ooops, I just hit [Ctrl]+[H] again! Must move my hand over to the mouse because that's more efficient than a keyboard shortcut, don'cha know.

Now, if I have a block of text highlighted and I select the Replace tool, do you think it's possible that I may want the replacement to be done within that block? Yeah, me too. The designers of the VBA IDE, misguided fools that they are, thought likewise and implemented it accordingly. But this tool doesn't. If I select Replace All it will replace all occurrences of the word in the tab, not just those in the highlighted block. If I just select Replace, it will replace the first instance that occurs after the block. So remember to put your cursor before, not on, the first word that you want selected, then repeat for each word in the block that you want to replace it in.

Ah, here's another block that I want to do a search in. [Ctrl]+[H]... oops, I mean, click pretty web-like icon that looks like the other pretty web like icon. Oh lookie, unlike what you get in Word (or most programs that I use) the dialog hasn't remembered the last search that I did, so I have to enter it in again.

Ah, a useful feature; when you hit [Return] after a function then it will automatically put the semi-colon after the line. Well done.

Oh, wait... but when you get to the end of a variable assignment line (like s = 'y') then the content assist dialog will appear. So better remember to hit that closing semi-colon then, because otherwise you'll get

Code: Select all

s = 'y'ABS(

since the content assist dialog will assume that you want the first function that it presents to you to be inserted when you press [Enter]. Always.

Speaking of which, the Content Assist dialog is another minimalist attempt. 10 years in the making, it hasn't quite reached the efficiency of the Intellisense system that MS had in VBA back in Office 97. Want to do an asciioutput? Fine and dandy, start typing Ascii, then press [Ctrl]+[Space] to bring it up (oooh, an actual keyboard shortcut!!! Slick!) and hey presto, you're told that the first argument is a string, and the second argument is a string, and there are more arguments which are probably strings as well. None of this mollycoddling of giving you descriptive names explaining (or at least hinting) what each argument is the way Intellisense does, this is a Content Assist dialog for Real Men (TM Iboglix Software Corp) where you get to manually look up what each argument is and what the function does. I say "manually" because (again, unlike the 15 year old Intellisense technology), you can't select the keyword and hit [F1] and have the relevant help page appear. Of course, if I can never remember which of the 7 separate help documents the relevant entry is in, how can I expect poor old Content Assist to?

You still can't Save As over a process that already exists; that was pretty much expected. Disappointing, but expected. At least unlike the Find dialog, the Save As one is populated with the existing name so that you can modify it instead of having to retype the whole thing. (Useful if you're saving with a similar name.)

I shall continue my exploration later, when I again get in some snooze time waiting for Performance Modeller to launch after (yet again) presenting me with the dialog asking me to me to open provagent.cogrcp_modeler with IBM Cognos RCP Application Updater. (Granted, I haven't rebooted since installation; I'm hoping that will go away after I do because the diesel loco that is Performance Modeller takes long enough to fire up as it is.)

[Andy Key]

Actually... according to Oxford it is correct for the US, just not for most of the rest of the world.

I'll blame Microsoft then. I changed my VM's regional settings and Office's language settings to US and Word still gave me red squigglies for Modeler.

Half the time I have to stop myself from typing "Performance Manager"...

I vote we rename it for them then, I keep calling it exactly the same thing!

[Alan Kirk]

Actually... according to Oxford it is correct for the US, just not for most of the rest of the world.

I'll blame Microsoft then. I changed my VM's regional settings and Office's language settings to US and Word still gave me red squigglies for Modeler.

If that was in Word, there's a gotcha; blocks of text have a language attached to them. You can see the language involved on the status bar at the bottom. So if you originally typed the text while the setting was English (Australian), changing your general language setting doesn't change the language of that particular paragraph. You'd need to do that via (in 2003; don't ask me about that P.O.S. 2007) Tools -> Language -> Set Language for that selection. If you don't, that paragraph will continue to be spell-checked using the Australian dictionary, not the US one.

Half the time I have to stop myself from typing "Performance Manager"...

I vote we rename it for them then, I keep calling it exactly the same thing!

I'd agree that "Performance" is not the name that comes to mind. With each passing use I come to understand more and more what The Cabin Boy was saying when he said:

Cognos TM1 Performance Modeler seems interesting though - anyone got an insider track as to whether it's worth the wait and what it looks like, because otherwise there seems to be precious little over 9.5.2.

You mean quite interesting in an 'OK, so I've gone to a web page and I've clicked an obscure button that launches a download that takes so long to load the splash screen let alone the full application that I might as well go and make a cup of tea' sort of way? Or the 'I've opened a rule and a cube view so I can see how my rule changes affect my cube, but now the panes of my single window are so small as to make them useless' way? Or the 'I want to see what the difference is between the rule in my Dev, UAT and Prod servers, what? I've got to go back to that web page and open up two more copies of NonPerformance Modeler as I can only see one server in an instance at a time?' way?

We still don't (as far as I can see; I have yet to install it) have a decent TI editor in which to write the code

Depends if you like the 'Advanced' rules editor and whether you like 'Performance' Modeler. I like neither. You may have guessed that already...

I'm getting the same impression, live and in 3D Technicolour. I don't know what we should call it, but for some reason the image of a giant, lumbering sloth wearing a coat with embroidered web icons comes to mind...

[Alan Kirk]

Odd, I can't seem to find "Applix.TM1.API.dll" in the bin directory in TM1 10.1 - has this been depreciated?

Edit: Even more strange - This does come with the installation but only exists in the TM1 Web Directory.

I don't think it actually matters (from a technical point of view) where it is, but I'd agree that (from a practical point of view) someone has screwed up and had it installed into the wrong place. The oh-so-useful .Net API .chm file appears to have vanished as well, with no "real documentation" replacement that I can see.

On the last point, I received advice from Support that the .chm will be reinstated in 10.2 and be made available as a download shortly.

No explanation of why the library has been shifted into the web folder though.

They also said that there are no plans to discontinue support for the .Net API.

[jim wood]

That's a bit of good news I guess.

[David Usherwood]

Has anybody looked at the new Synchronized TI keyword? I'm finding the documentation less than transparent eg:

TurboIntegrator process may make any number of calls to synchronized(), with
any number of lock objects. Serializing is effective from the time synchronized() is
called, until the containing transaction completes.
For example, if synchronized() is called from a subprocess (Ps) of master process
(Pm) or master chore (Cm), the Lock Object is "released" when Pm or Cm
completes. The exception is that a SaveDataAll (SDA) prematurely "ends" a
transaction mid-process execution; this applies to Lock Objects as well.

So is it just a flag saying 'stay off the cubes I'm updating till I finish'?

Also (think this might have been covered in a post) 'break' is a reserved keyword, but not documented in the Reference Guide. Anybody played with it?

[David Usherwood]

Found the 'break' thread...
http://www.tm1forum.com/viewtopic.php?p=24933

[nick10598109]

But other than that it ran pretty smoothly. I installed the whole smash, and it took about 26 minutes all up. I had the same experience as Steve with item 167; whatever it is it took the bulk of the time though mine was 167 out of... 288 I think it was. The end dialog reported that the Configuration Tool completed with warnings and errors but there was nothing in the log about what those might have been.

The place where it looks 'stuck' is caused by the large number of documentation files that it is copying over and that it is installing all languages. This is expected behavior for the 10.1 installer.

What were the warnings and errors in the logs? Using Windows Server 2003? I'm guessing it says that the maps folder is missing. I've seen that the new Cognos installer expects a maps folder but can't find it, since it is not needed for TM1. You can ignore these errors.