Captain Oblivious

1 - You are right. It is a little complex to input so many parameters to setup a development environment for Notes8.0. But it is because that we only use Eclipse as the development(no development toolkit installed). If you hve Expeditor Toolkit installed on Eclipse. The process of setuping dev enrionment will be much easier.

2 - Excellent rant, Captain. It's especially funny that "Composite Applications for People" comes immediately after "Captain Oblivious" in my Bloglines reader, so I read your rant before the actual article-- and subsequently, couldn't read the Composite Apps article with a straight face.

Yes, there is definitely a "rite of initiation" involved with a lot of this stuff. At least for a developer like myself, examples of this are all over the Admin help (like I wanted to set up the DB2 backend on my Domino 8 test server, but got a little scared reading through all the steps).

And of course, the classic case is Websphere Portal install -- Volker's always been amused by that one. Note that I've heard that the new Portal install is supposed to be much better...

But your point still stands.

3 - No, you're not alone. The move to Eclipse was one of my reasons for getting out of Notes development. I've tried to use Eclipse off and on for a very long time, and each time end up so frustrated I just walk away from it (after cursing long and loud).

The work Maureen Leland showed of Domino Designer in Eclipse had me excited, but then I realized it's going to be years before that is released. The cynic in me says that even when it is available, it's still only going to be an 80% solution because IBM has never done anything 100%.

I'm glad IBM has people like Bob Balaban, Andre Guirard and Maureen Leland working on the development toolset problem. They're a brilliant lot, and hopefully they can come up with something that will make me want to use my favorite platform. Until that happens I'll have to settle for a toolset that doesn't make me want to cry daily.

I still consider Domino to be the best collaboration platform available, but the current development toolset does not fit my needs or wants and the direction I see IBM taking it doesn't match where I want to go.

4 - I think you're missing the point, to be honest.

Launching Notes -- essentially adding the Lotus Notes Plug In to an existing Eclipse IDE is not a design goal or requirement.

To most (I'm guessing over 99%) of people, there will be no separation in their minds of the Notes client from the Eclipse based framework it will be living in for those whose companies go that route.

Even fairly advanced developers don't need to launch the product from in Eclipse like that unless they're also doing some fairly significant other Eclipse based work in which case this isn't really out of the realm of reasonable (though it is more complex than it should be).

For the ENTIRE existing Notes/Domino development community, nothing at all HAS TO change. What the Eclipse framework DOES give us as advanced developers with willingness and skills in other languages, is to break down the wall between the supplied UI tools and the UI features we'd like to be able to add.

In that case, the job of the Eclipse framework is to provide the same set of functionality that is provided by the Microsoft Foundation Classes for Win32 based development or the OSX ('name of the day') UI standards tools -- or gnome or kde if that's your cup of tea.

Can you suggest ANOTHER way to do that, which will be cross platform?

5 - @4 - Whether there's an alternative for the goal isn't relevant to Rob's point: that the method of achieving the goal is so ridiculously byzantine as to be insulting.

And is the goal itself all that important? I've had some lengthy discussions with the composite apps team, and I have yet to see a composite app that I couldn't build in the Notes client using COM. I find myself questioning whether there will ever be one. In which case, do we really NEED all this infrastructure? What did we give up to get it?

IBM has taken a huge gamble here. Think of what capabilities we might have seen in the C++ version of the client had it not been for the Eclipse-wrapping effort. What solutions might have been possible with a few minor items in the native client that IBM didn't provide under the auspices that "you can do that with a composite app?"

I'm sure they'll make it better with time, but Rob's post nails exactly the point that I've been making as part of the DP group for months: Composite applications are a way to get Eclipse developers build Notes applications, not a way to get Notes developers building Eclipse applications.

6 - The point to me seems to be: give us a one or two button click process to do this stuff, not 2 pages of instructions.

If IBM knows how to set it up, it would be great if they could spend an extra day or two of developer time to put some kind of install routine together.

I understand that all this stuff is infinitely configurable, but let the people with the weird configuration option needs tweak things after the install, so the rest of us get an easy initial install.

7 - @6, @7 -- Argh.

The minute you move from NOTES APP development to Eclipse Plug In development you leave the world of "give me two buttons" purposely behind.

That's the point of it. You're talking development which is more like programming for a complete application than building a "notes app" which is more like (and grew from) building a fancy spreadsheet or access database.

Nathan, you can't honestly tell me that you're frustrated at times by what the Notes UI is capable of handling.

If you want to break those barriers, you're into complex programming. That's the way it is.

If you are already doing big time Eclipse based work, you MIGHT want to add in the Notes stuff. This is how. It's hard.

Me? I prefer to just have a distinct eclipse environment for that work, and the IBM Installed Eclipse desktop for the Notes stuff.

8 - Wow, I must have seriously overestimated the Eclipse development environment, if someone can't develop a one or two click setup routine to get configured to "launch Notes8 from within your eclipse IDE". Which is what the article tells you how to do.

Every other Eclipse plugin I've ever installed has been pretty automatic. And many of them are extremely complex.

9 - And to be clear, Rob specifically said "throw a simple installation routine together", and I talked about setting up DB2 backend integration and setting up WebSphere Portal (comment #2). I agree that the *development* is going to have some degree of complexity. Just give us an easy setup.

10 - Julian,

I absolutely agree that the installation routine is overly complicated -- even said so. I just don't see where it rises to the level of vitriol. I frankly am a bit surprised its offered at all.

Keep in mind, The framework used for the Notes client (as planned for IWCT-RE) is based on, but not within the same fork as, the current open source Eclipse project.

I think the abject failure of Workplace has led to some lessoning of the need to make that framework quite so distinct, but that's pure speculation on my part.

11 - @Andrew: Fair enough. I know that the new framework tends to matter less to someone like me -- I'm a Notes guy who suddenly has all this extra Eclipse stuff to play with, but it's already built-in (per your comment #4).

It'll be interesting to see how well it goes the other way, if the Eclipse folks start trying to play with Notes.

12 - @7 - Of course I'm frustrated by the limitations of the Notes client UI. I've make a career over the last couple of years out of getting around those.

But tell me, what would be more empowering: 1) putting the limited Notes client inside a wrapper environment with fewer/different limitations? or 2) getting rid of those limitations within the Notes client itself?

After all, it's not like the constraints of the Notes 7 client where some sort of metaphysical given that IBM had no control over.

Could the same set of resources that went into turning Notes into an Expeditor application been used to solve some of those other problems? Could they have given us, instead, programmable embedded views? (Oh, wait, I already did that.) Or perhaps cross-NSF Lotuscript UI objects? (Oh, wait... I did that too.) What if all that investment had been made, instead, in building better out-of-the-box templates for Notes 8?

That's the gamble IBM made: that the value of having Notes be a composite application container is greater than the value of having Notes itself be a better development environment. Whether or not this strategy will pay off WILL NOT BE VISIBLE IN NOTES 8. And it's only going to pay off if the developer base for Notes is not put off by the implicit statement of "go over here and learn Eclipse if you want to do anything decent."

A dev environment is only as good as its toolkits, and right now, Expeditor doesn't have much of one from the perspective of the Notes developer.

13 - It sounds to me like you have an alternative plan, Nathan, that would allow a truly flexible fully event driven GUI for the Notes client that would would be cross platform, but not built on the Eclipse framework and would be easier to integrate with and build for.

I'm excited to see this product you have in mind. Can you share it?

14 - @13 - Are you being sarcastic, Andrew?

A flexible event-driven, cross-platform GUI that's not built on Eclipse and is easy to build for... yes I have a product in mind.

Notes 7.

Why are we debating this?

Does Expeditor make it better? Well, obviously, yes. But the ways in which it makes it better are relatively inaccessible to the existing Notes developer base. Surely you don't consider that a point of contention, do you? Do you think it's as easy to build a composite app as it is to build a Notes app?

Do you consider the strategy of turning Notes into an Expeditor application to be risk-free? Do you think that if IBM had retasked those resources responsible for making Notes an Expeditor app, they would not have been able to, instead, make a Notes a better app all by itself? Does Notes 7 represent the pinnacle of what's possible in cross-platform C++?

By the way, if you want me to name a different flexible event-driven, cross-platform GUI that's not based on Eclipse and is easy to build for, how about Firefox?

15 - I think "expeditor" as its now called, may be the single most important thing that IBM has done for the Lotus Notes product focused ISVs since they acquired Lotus.

I remember when Lotusscript first made it into the native client in version 4, how even the smallest scripted applications took high end consulting and many hours to build and test.

I remember when if you knew how @DbLookup and @DbColumn worked, you were ahead of 90% of the "Lotus" develpment community.

"Expeditor" has opened the Notes Client development world to higher end techniques and results in a way that for the first time may lead to actual high end third party applications that LOOK polished and high end to an end user base that is used to seeing full blown GUI applications.

If that means only 2% of developers figure out how to take advantage of that in the first 3 years of the products life, it will still be a success.

What's needed next is the same kind of opening for the Web side.

When you speak of "retasking" those resources into the Notes client, you need to go back to 2003 and remember that for all intent and purpose development resource to the pure Notes client were DEAD and GONE. What we now call "Expeditor" was originally "IBM Workplace Client Extentions - Rich Edition" and the fact that Notes ran inside it as described "Sharing the glass" without real integration was considered by many to be a win and a relief for the Notes community.

What the team in Westford (mostly) did was to take advantage of a well planned architecture that had always been built for cross platform compatibility and not only rapidly exceed the expectation for their ability to integrate. They were so overwhelmingly successful with it that they rapidly surpassed what the IWCT team was able to do with their own platform, and were soon setting the specifications instead of following them. By the time "Workplace" as a specific product suite collapsed, it had become clear that the Notes client running within that IWCT-RE framework was by far WAY more powerful than any "Workplace" product based on J2EE was every going to be.

You can't take the positive outcome we have now, and re-render it outside the historical terms that created it. That would be like saying "Imagine how much better our new pebble bed reactors would now be if we'd skipped that whole generation of melt-down prone 'control rod' technology based power plants from the 70's." Its meaningless speculation because it could not have happened that way.

16 - @15 - Wow, Andrew. Consider me lost. I have no idea what you're talking about.

17 - @15 - Actually, let me clarify.

The fact that you expect me to understand your pebble-bed reactor analogy illustrates my point about accessibility more perfectly than I could ever do on my own.

18 -
LOL -- In more plain terms, all metaphor removed:

It is fallacious to suggest that the resources spent on getting to the "Expeditor" version of Notes could have been instead spent on the Notes client directly without insisting on an Eclipse approach because the real alternative was to cease all spending on the Notes client.

The work done on Expeditor was not originally meant to be of benefit for the Notes client, it was for the replacement product "Workpace" (Yes Ed, if you're out there, I know you won't like it framed that way). Notes was to take advantage of the new framework "where and when it could", and in the initial release would have only the most minimal interconnectivity with the rest of the framework based tools. This was called "On the Glass" integration -- they would share the same screen without causing a fire, but little else.

The evolution of "IWCT-RE" with the Notes plug-in into the "Lotus Notes version 8" product with "Expeditor" a positive outcome to a "make the best" initial approach.

My further point is that the complexity being ranted out by the good Captain Oblivious, is one which need not exist for the vast and overwhelming majority of Lotus ISVs, let alone developers of most Notes applications. Rather, it is a way for people doing completely different Eclipse based work who insist on having a single Eclipse desktop (which is itself not necessary) to do so.

@14 -- if you think you have serious cross platform UI control in Notes 7 that even remotely represents what can be done in the Eclipse framework you're fooling yourself. Especially if you take out of your tool bin anything related to COM, Active-X, or other Win32 technologies since the big push is to FULLY support the Apple and Linux desktop community (of which the latter is still a very tiny minority which doesn't seem to be growing a lot).

----------------
There is a real problem here, however. The problem is that a smart guy like C.O. doesn't have enough real, valid, information to tell the difference yet between what he's going to need to learn to do and what wild alternative things he may do if he so desires, which means that the complexity, capability, advantages, and roadmap for the Notes 8 development is very poorly communicated so far.

19 - @18 - Eclipse version of Notes vs. no more Notes client at all? That's a false dichotomy. There was and is no TECHNICAL reason why we couldn't see massive improvements to the native C++ client, just as we had for the previous 7 versions. The move to exploit the previous cross-platform work done in Eclipse was a BUSINESS decision.

You keep saying "realistically, this was the only way it was going to happen." Fine. But that's the same argument we hear time and again from, say, CIOs switching from Notes mail to Exchange. Somehow, decisions are just MADE without consideration of the possibilities, and we're all just compelled to go with that. If that's how you want to look at the evolution of this platform, I'm certainly not going to convince you otherwise.

But regardless of the motive, the fact remains that IBM (and you) are now positioning Eclipse as the necessary platform for "serious" GUI development involving Notes. It's supposed to be the solution to all the limitations we faced before.

The pitch is: here's how we solved the prior limitations of the platform -- by wrapping it in a bigger, more powerful platform.

Well, okay. That's an answer. Is it a good one? Is it a better one than "okay, we've bundled Mozilla as the native browser application in the Notes client and given you these 6 Lotuscript classes for dealing with HTTP content rendering in the client?"

I dunno. It's certainly valid to question it, though.

Certainly taking a long-time Notes expert like Rob and putting the Eclipse dev architecture in front of him and having him go "WTF?" is indicative of the solution needing more work.

20 - Well, you keep suggesting that just 'adding on to the C++ client' would be your way to go.

First, unless I'm radically mistaken, it isn't c++, its just C.

Second, no matter how much they do in that manner, you'll still always be stuck with the dialog boxes and menus they are able to provide to you with your more abstract tools. Sure, they could do a lot more in that area -- but never enough to give you real control to make your OWN window and control it your way. Not within that model and especially not if they wanted to keep it cross platform.

That's the whole point of the Eclipse project itself. To provide an open alternative to the Win32 foundation classes, (or now the .NET framework CLR), the KDE/GNOME desktop design requirements, the OSX (is it still Carbon? I don't follow mac) standards for UI, etc., etc.

A full blown UI framework is a BIG DEAL. They've ported Notes to run in one. That means you can now use the full power of that framework.

Yes, it was a business decision. EVERYTHING is a business decision. The alternative at the time was not so much to "get rid" of the Notes client -- of course that wouldn't happen. It would, however, have gone even further into "Maintenance Mode Hell" and eventually gone the way of AMI PRO and LOTUS 123. The big money was on Workplace, and Notes got to play in that sandbox because there was too much pressure from the install base not to.

Its a good thing, too. Had that not happened -- had they not started playing together, then when Workplace failed there would have been no really viable way back.

I am pleased to note, that in November of 2004 when I gave the first Non-IBM overview of "Notes Next within IWCT-RE" in London one of the key points I made was that "Workplace will live or die on its own merits, but either way the Notes Client gets a lot of new functionality as a result, so that part will continue to grow." If you recall, at the time the big worry was if Notes would continue at all long term in a world of Workplace.

At the end of it all, Notes has actually gotten a solid and complex framework to run in that is much more accessible to program in than pure C -- especially if your C code is required to do UI work for each platform.

I'm waiting to hear your alternative. All I've heard in that direction so far, is "bundle Mozilla as the native browser application in the Notes client and given you these 6 Lotuscript classes for dealing with HTTP content rendering in the client"

That's an interesting and possibly valid approach -- but totally small scale thinking. The new framework would (A) let you do just that if you wanted and were up to it, and (B) give Lotus the place from which to begin doing that kind of work. You've seen the things Maureen is in her plans for direction and you know that.

I know of no platform other than Eclipse that provides a UI framework acceptable to users that you can realistically deploy to Mac and Win32, Gnome and KDE to run your own apps in.

Again, lets here your big plan.

21 - I have been corrected several times by folks from Westford that the basic Notes client is not C but C++. Perhaps they were mistaken.

"Sure, they could do a lot more in that area -- but never enough to give you real control to make your OWN window and control it your way. Not within that model and especially not if they wanted to keep it cross platform."

Clearly, Andrew, you don't read my blog. I'll assume you didn't attend my Lotusphere session either.

"They've ported Notes to run in one. That means you can now use the full power of that framework."

Huh? No it doesn't. Maybe it means SOMEBODY can. But I, as a Notes developer, have basically ZERO access to that framework. It's even less accessible to me than, say, a print dialog box is in Notes 7.

Show me the Notes 8 code where I can manipulate the menu structure from my NSF in a way I couldn't in Notes 7? Where's the @function to add something to the RSS reader? How about Lotuscript to format a cell in the productivity editor?

This is a one way street. Expeditor can program behavior into the Notes client -- not vice versa. It's targeted at someone other than Rob... or me, frankly.

"I know of no platform other than Eclipse that provides a UI framework acceptable to users that you can realistically deploy to Mac and Win32, Gnome and KDE to run your own apps in."

The web browser.

22 - The web browser is f'ing terrible place to work, to write code for, or to suffer through doing anything for real.

Sorry, whole architectures keep being invented and failing just to make the web browser somewhat workable. The reason people keep trying is because it its base, it blows chunks.

As far as development capabilities, the issue is that this time its backwards and that's why your underwear is in a twist. In all previous Notes revs, you end up upgrading the client side because you want to take advantage of new development features. Lotuscript, frames, @dblookup, embedded views, whatever it was in that version. This time around, the client came first, so yeah, the development tools to take advantage of the new stuff aren't there yet. Do you doubt they're on the way? You've seen better.

This time, for a variety of reasons, the client side had to be first to define how the framework would end up getting used, and it wasn't until that solidified that the development tools could get seriously worked on. Holding the entire release for a year while the dev tools catch up would be serious folly.

23 - Ah, so you're saying they haven't actually solved the problems yet -- just given us a framework for eventually doing so.

Now where did I read something like that?

"Whether or not this strategy will pay off WILL NOT BE VISIBLE IN NOTES 8. And it's only going to pay off if the developer base for Notes is not put off by the implicit statement of "go over here and learn Eclipse if you want to do anything decent.""

24 - Only if you're thinking about the average it developer.

Seriously big time ISV's will start very soon putting out top quality ui driven stuff that interfaces with Notes for the first time ever.

You're going to see really classy looking Notes stuff for sale in shrink wrapped kinds of deployment models that actually look good enough to justify their cost for the first time.

That will drive an improved Notes environment, and yeah, the rest of us will have to work to catch up even as the tools get better and bring that stuff closer to end "user-developer" types.

The bar is being raised in terms of what can be done, and in terms of how hard it is to do. Just line in version 3, 4, 4.5, 6, and a tiny bit in 7.

Remember the notes "developers" that never transitioned to lotusscript?

25 - @24 - That's a market prediction that I can't make any claims about.

But you've agreed with me, then, that this composite application effort is not for the existing Notes developer base, which is pretty much all "average it developer(s)." It's for "seriously big time ISV's."

I'm pretty sure that's what I was saying. Lemme check...

Composite applications are a way to get Eclipse developers build Notes applications, not a way to get Notes developers building Eclipse applications.

....yep... looks like that's what I was saying.

26 - Of course it starts with teh big guys, they can afford to hire the expertise you and I will have to learn, and many will never both with.

Shocking.

27 - Now, this is the last type of post I would usually comment on... em.. as you all lost me ages ago, but Rob and I share something in common here. I am completely and totally frustrated with the expectations that IBM has in relation to the interest of the IBM customers/administrators/developers as a whole.
IBM seems to expect that we will kill ourselves with
1. Brutal (i mean woefully terrible) documentation
2. Overly complex procedures
3. Massive integration issues
4. Version compatability rules (i.e. - to get product x running you MUST use version 4.0.1.2 of product y... not simply version 4.0.1.2 or higher... that EXACT version).
and still champion the products as a whole. This is a critical time for Lotus Notes - 8 is almost here...
This shit needs to be EASY to get people to USE it. And believe me, Admins and Developers need to find it easy to set up.. .or they will lose interest quickly enough (the MS suite is easier to get up and running... and there is a hell of a lot more support out there).
Im not asking for any magic buttons to get complex things running - but c'mon... the basics should be relatively simple. IMO (and believe me.. I'm not looking to be flamed here..) Notes developers will only look at plugin/eclipse development if its relatively easy to get started and get their head around it... They are used to RAD apps and FILE/Database/New...
Im not saying they shouldnt have to change or adapt... but there should be
1. excellent documentation
2. excellent demonstrations online
3. Wizards to just get up and running..
I think we need to sell this to the developers... Not the genius developers in the world like the guys above, but the developers that do this from 9-5...

Another question... If the Millions of Dollars that went into Websphere or Portal (or dare I say it Workplace) went into Lotus instead over the past 8 years... what would we have now?
sorry guys... rant over (a very very frustrated Notes evangelist ..currently doubting his faith)

28 - @27 - But then why would we need ISSL?

29 - Note to new comment readers: Rob has a follow-up blog entry:

{ Link }

30 - I have the same reaction as Rob - I read the Composite Applications blog because I think I should be getting on top of this stuff, but it just gets scarier and scarier.

I too am frustrated at IBM's approach to development tools. Just last week, I attempted to get the Workplace Forms server API installed on my PC to do some development.

After running the installer:
Step 1 - copy files from folder x to folder y
step 2 - manually update .ini file (including very long path strings where the current version number had to be inserted - doco just had .../x/..... and you had to guess what it meant)
step 3 - manually create prefs file (ditto file paths)

and so on. Stuff that should be in the installer.

We Notes developers have probably been spoiled in the past, but I see too many of this type install from IBM. Tools & environments that are difficult to set up, not version compatable, etc. Easy to use tools will win out over more powerful but difficult tools in 9 out of 10 instances.

31 - Getting back to the point, re configuration in Eclipse, I agree that this stuff is grim.

But just code some Websphere applications for a few months, writing Ant build files, locating esoteric Java driver packages and the like. Two pages of Byzantine instructions are like a walk in the park after that...

32 - Getting back to the point, re configuration in Eclipse, I agree that this stuff is grim.<br /><br />But just code some Websphere applications for a few months, writing Ant build files, locating esoteric Java driver packages and the like. Two pages of Byzantine instructions are like a walk in the park after that...

33 - Not much to add, except thank you for bringing this up.
If this is tough for those with 10-15 years experience, what is a first timer supposed to do?

34 - I think that Nathan exagerates quite a bit about the general "inaccesability" of eclipse for notes developers. Notes development evolved over time, too. I find the codebase in todays survived notes-shops quite different than it was 4 years ago. They tend to use more classes in LotusScript, XML sometimes and I don't learn the new RichText classes in a hour (need more time).
To get the basics of Eclipse plug-in development takes about a week (if you have the right book(s)).

35 - Yes, the setup for Eclipse environment for Notes 8 can be painful with the betas. The following is direct from the development team "The items that are called out in the blog are exactly the set of items that the [upcoming] Expeditor Toolkit simplifies with respect to configuration and launch. The Expeditor toolkit will provide a much better experience for developers than the current set of instructions provided with Notes beta or Sametime."

- Happy coding.

36 - And, to add to Mark's comment (35), Expeditor Toolkit will not be the end of the story, either.

While Notes 8.0 offers great new functionality and has some tremendous flexibility for developers to extend and customize, some will still be disappointed by the existence of some holes in the front-end scripting situation, and many will probably complain about the necessity of writing Java plug-ins to get certain things done.

I am working with several teams in the Notes Client Development organization to address these issues for NMFR (next major feature release).

Please remember that we can't do everything in a single release (the fact that we got as far as we did with 8.0 frankly amazes me at times, it was a huge gamble).

So, stay tuned, and keep that feedback coming. Thanks!

37 - IBM has already lost many (most?) Notes developers. I have had recruiters call me and say that they didn't think there were any Notes developers left. There is virtually no demand for this skill in the Boston area, but when a company needs one, they need one really bad. Sadly, only short term contracts seem to be available.
Most Notes developers are non-traditional programmers who picked up the skill because of the ease with which you could deliver a secure, networked application. As I try to move to Java/Eclipse I am realizing that the Sharepoint suite has much more in common with Notes as I know it than ND8 and Websphere do.
It has been a long bumpy ride the past 7 or so years and I find myself wishing that I had just moved to another platform instead of investing this time in Notes development. Perhaps ND8 will improve the situation, but I'm afraid that it will take years for the transition to reinvigorate the Notes job market.
Whew - glad to get that off my chest.