Fear and Loathing in Lotus Hannover
Okay, maybe my title's a bit melodramatic, but I'm worried about the UI quasi-revolution going on a Lotus right now. We see screenshots of Sametime pre-releases at Ed's blog. Mary Beth has a new blog soliciting feedback about the UI for Hannover. It's fantastic that Lotus is digging into this much-needed effort and generating excitement about it among the community.
I'm honestly worried, though, that they're missing the forest for the trees on this stuff.
It's long-standing lore that the GUI revolution over the CLI was that, since graphical interfaces so limited the choice of actions by the user, the computer became comprehensible. The blinking CLI prompt offers the user an infinity of choices. The GUI, even complex ones, constrain the users to a few initial gestures with the mouse, each with fairly intuitable meaning. We all know this.
I believe that one of the great strengths of the browser-as-application-platform that has been so strong over the last decade is less about ubiquitous access and more about again constraining user actions. Generally, you don't typically have to train someone how to use a web site. It's back, forward, go and refresh. Even modern Ajax applications focus on speed of information delivery and the application reacting to the user, not truly new gestures from the user.
Witness the success of the iPod. One of the things that makes it successful is it's radical simplicity. One small LCD and a wheel with 5 buttons. That's all you need. The interface works in part because what an iPod does really isn't all that complicated, but also because there's no way to get lost using one. The user is so constrained that there's simply no way to do something wrong.
<RANT>
Rich client software, in the meantime, has run rapidly in the opposite direction. In the rush to implement features, we've ended up thrusting a near-infinite set of options back at the user. GUI design 101 says that users get confused when presented with more than 7 options in a list. The context (right-click) menu in my Notes inbox has 27 choices! TWENTY SEVEN!!!! It fills the screen in 800x600 resolution! The inbox has 9 action buttons across the top and 18 items in the Actions menu! How do you digest that? I've been writing software for this platform for almost 15 years, and I get lost in my own email!
If Lotus wants to overhaul the user experience for Hannover, then the design center has to be radical simplification. Mary Beth wants to know if it's okay to remove certain menu items. Screw that! I want the option to turn off the entire menu bar! And I want the option, as an administrator, to set that via a policy for my users. (Seriously, do we need a menu bar? It's almost 100% redundant to SmartIcons, so pick one or the other.)
Let me put this another way: Lotus is asking the wrong questions. Static interfaces are yesterday's news. I want the ability to turn EVERYTHING on and off. I want the ability to alter every single menu item on the screen. I want control of everything in the bookmark bar. I want to be able to set mail preferences that completely hide entire functionality sets (I don't use To-dos in Notes. I don't want to use To-dos in Notes. Why is this persistent throughout the interface, then?)
Don't ask me if I need a menu item. Generalize. Give me the ability to customize my menus and my UI preferences top to bottom. Then, when you have that, make it policy controllable, so I can impact the user experience for 10000 people at once.
Don't tell me "it's Eclipse... you'll be able to write code for this." I'm the NTF in OpenNTF -- no duh! The point is that this should be controllable by end-users and administrators, not code-jockeys. When a 200 person shop on an Express license decides to implement Notes mail, their 2 IT guys should be able to say "let's turn on full-text searching for these users, and not for these users." "Let's roll full C&S capabilities to these 20 people, and just offer mail to the remaining 180." "Let's enable IM for these 50 users, and not these 150." When a feature is off, it shouldn't be greyed out, or just have blank configuration. It should be INVISIBLE. Because if the service isn't offered to that user, they're just forced to shift through more chaff.
Lotus, I know why you present things the way you do. You have a great product with an incredibly powerful set of features. Admit it... you want to brag about it. I would too. Actually, it's not even my product and I brag about it all the time. But it's a disservice to the user. It's messy and therefore confusing and therefore frustrating. And that's the reason you take a beating on Wikipedia and Slashdot.
Try an exercise: How many different mechanisms in Notes 7 exist to reply to an email? Notes 7 gives you 8 different options in the action bar for replies. 3 more in the context menu. 8 more in the Actions menu. Another in the Create menu. The fastest one still takes 3 gestures! (Note: one of the very first features of the OpenNTF Mail template was to allow for a configurable default reply so we could shorten that to a single click. Of all the things that Lotus did crib from OpenNTF Mail, that should have been the very first.)
If you had mail on an iPod, how many interface vectors do you think there would be for replying? There'd be ONE, of course, and it would be the same whether you were looking at a list of mails or reading an individual message. How many ways are there to reply to a message in Gmail? Two. And one of them is only visible if you click on "more options." Check out Hannover and see if you think it's any simpler than today.
Sametime 7.5 is another product I'm scared about. The beauty of instant messaging is it's simplicity. You see a contact. You double click. You type. You hit enter. It's compact, it's simple. I rolled out the instant messaging capabilities to people in my office with ZERO training, and everyone gets it. I've never had one tech support call.
Now look at Sametime 7.5. Rich-text, pictures, avatars, bright colors -- it looks like someone spilled a bag of Skittles on the screen. And a single conversation consumes almost twice the pixels as the current version of ST, and nearly 3 times what I see in Trillian (which is a great example of letting me turn things on and off.) Yes, the new whiteboarding capabilities are great. I'd love being able to have multiple sessions and gateways to public IM services and all the other great new tools in 7.5. But probably won't implement it, because there will just be too much crap on the screen when I do what I'll do 99.9% of the time with Sametime: have a quick chat with the same 5 people I chat with on a daily basis.
Lotus, please don't lose the plot on this. You need RADICAL simplification. And because you can't start stripping features, you need to do the next best thing -- let me as a user or me as an administrator decide what's important for MY ENVIRONMENT. No two Domino implementations are alike. Stop browbeating your users with all the incredible things they could be doing, and just let them get on with what they are doing. You have a great opportunity to leave the competition in the dust here. You've avoided big moves for long enough -- throw out the baby, the bathwater and the whole darned tub!
</RANT>
Let me reiterate that Sametime 7.5 and Hannover are very exciting releases. It's just not far enough. The whole community needs to encourage IBM to get serious about rethinking how they represent the Notes product, because if this is the future of activity-centric computing, it's gotta get smarter and simpler.
UPDATE: I think Mary Beth might be my new favorite IBMer. (Sorry, Ed.) Say hello to the new, responsive IBM!
Comments
Posted by Carl At 10:32:27 AM On 04/20/2006 | - Website - |
I think the point is that the LAST thing Notes needs is MORE flexibility (different flexibility, changes to flexibility, sure). What it really needs is USABILITY.
Posted by Samuel Allen At 02:32:11 PM On 04/24/2006 | - Website - |
I was the designer. :)
Posted by Mary Beth Raven At 07:28:33 PM On 04/20/2006 | - Website - |
Posted by Sean Burgess At 11:22:13 AM On 04/21/2006 | - Website - |
Second. I hear you. You want the ability to customize all the menus, all the preferences and control it by policy.
Posted by Mary Beth Raven At 11:01:52 AM On 04/20/2006 | - Website - |
But I dig the one-ring-to-rule-them-all approach on the physical unit itself.
Anyway, if you don't like me relating to iPod, then focus on relating to Gmail.
Posted by Nathan T. Freeman At 11:34:24 AM On 04/21/2006 | - Website - |
Posted by Bob Balfe At 04:58:42 PM On 04/22/2006 | - Website - |
My sentiments precisely
http://www-10.lotus.com/ldd/nd6forum.nsf/55c38d716d632d9b8525689b005ba1c0/c8dd8acef57dd3c585256f12002a0198?OpenDocument
@Samuel
I don't agree, flexibility is very important the more the better. It is up to the application developers to make the applications more usable, which they can if they have more flexibility. Remember Mail is just an application in the Notes environment.
Posted by AJP At 05:10:44 PM On 04/24/2006 | - Website - |
I agree wholeheartedly with Nathan's assessment and the others who echo his sentiment. We need USABILITY, and we are getting rather desperate.
For those who say "wait for Eclipse", I really don't think I should have to become an Eclipse config file super ninja just to give my users the customized set of features they want. Besides, do you have any idea how much time it takes to hand tune a mail template every time a new one is released? Right now it's about 2 - 3 hours for me to go through and rip out or hide all the extra stuff that is irrelevant for my environment. Isn't that what policies are supposed to be for?
Posted by Charles Robinson At 10:04:55 AM On 04/25/2006 | - Website - |
Dear NanTFan,
I'm going to say your passionate plea is about 80% convincing. I do like the idea of usability, and being able to turn on and off complexity.
Traditionally, IBM/Lotus has left the choices in the hands of the users and has only slowly been giving control to the admins. (There is a long list, here, from approving ID name changes to mail configuration.) It's a mostly good choice. Except, when the admin wants to simplify the features to simplify their own work load.
How many sites use their free Sametime IM integration? How many sites configure their servers for true shared calendars? How many sites allow DAMO or IMAP4 (lots of cellular phones can use IMAP4) or even OpenNTF mail alternatives?
My numbers say that about 10-15% of the Domino installations actually take advantage of Domino features which would dramatically benefit the users.
So, I'm 80% in favor of your vision, and 20% suspicious that many admins would lock-down Notes into an Outlook Express/Pine/YahooMail simple interface and wash their hands of any feature complexity.
Posted by Jack Dausman At 02:28:13 PM On 04/26/2006 | - Website - |
Another approach to GUI levels is a "basic" vs "advanced" toggle, the most classic example I can think of is Winzip. The user decides which version they are more comfortable working with. When they are in basic and are finding themselves too limited in terms of features, they switch to advanced.
As far as turning settings on and off via policy, I'd love that. I know a lot of devs/admins (specifically in the Health Care industry) that would love the ability to disallow any custom toolbar buttons.
I also don't like to see anything in a context-sensitive menu unless it's really context-sensitive. "Find" for example, in the right-click of an email memo leads me to believe it will just search in that memo. However, it's just a shortcut to the "Find" at the top of the view.
Still, though. I do love my Notes Client.
Posted by Jess Stratton At 09:24:05 AM On 04/21/2006 | - Website - |
I think it's a little unfair to compare Notes to an iPod just as it's unfair to compare Notes with Outlook. I think it's much easier to create a simple interface when you are trying to perform a specific task like read and write email or listen to music. The same thing is true when I am trying to write an application. The simpler the application, the easier it is for me to keep the application interface uncluttered. It's once the forms become relatively complex that I find my best intentions going all to hell.
I think it would be quite a gamble for Lotus to completely change the interface akin to what they did with Improv. I thought what they did with that product was revolutionary, but it was accepted by the user community and ultimately died an untimely death. I don't think that it would happen with Notes, but it's got to be the fear in the back of the developers mind.
Sean---
Posted by Sean Burgess At 12:29:21 PM On 04/20/2006 | - Website - |
Let me put it another way: I'm working on a MAJOR interface overhaul for an extremely complex product right now (medical practice management) and I have an iPod next to my monitor as my constant reminder that the most important thing I can do is to take stuff away and hide the complexity of what, underneath, is a mind-bogglingly complex system (accounting and clinical records rolled into one!)
@4 - Mary Beth, I'm delighted to hear that. If you can keep that sense of simplicity as you work on Hannover, then what I'm asking for here can be achieved. The reasons I wrote all this are a) because the screenies I've seen of Hannover thus far remain FAR too involved; and b) because I honestly believe that the posts on your blog at this point aren't reflective of looking in the right place.
That being said, it's perfectly possible that you guys are moving strongly in the direction I'm describing, but aren't telling anyone yet. I really have no way of knowing. (Speaking of which, how could you NOT invite someone from OpenNTF to the design review team!??!)
Posted by Nathan T. Freeman At 03:28:01 AM On 04/21/2006 | - Website - |
OK, I'm a lazy developer - I don't user reader/author fields except in some very rare cases, my bad. I use the GUI to manage what people can do and when. By making it impossible for a user to switch a doc into edit mode when they're not supposed to edit, I protect the doc. Can't write agents? Can't force date into the doc. Create a SmartIcon and I'm screwed (how glad am I that most people are too busy or uninterested to play).
I very much agree with your comments - less can be better and let me figure out how much is enough and let me control it from a central location. I spend most of my designer time making it hard for people to do the wrong thing - the less they have think about what to do, the better.
FWIW - <Apple Rant>
I don't find the iPod interface all that great - one big magic button for doing everything...which sequence got me to the 'fast forward through the talking to get to the music on this podcast'? Some days I feel like I'm playing Zork... And trying to get my iPod/iTunes setup working the way I want - grrrrrrrrrr. And don't even talk about what happens when you decide to move your library from one disk to another. And I can't delete stuff without connecting to a computer????
</Apple Rant>
Great post, can't wait to see how RNextNext turns out.
Doug
Posted by Doug Finner At 10:01:36 AM On 04/21/2006 | - Website - |
I've written about 7 different replies to this, deleted the comments, and started again.
I have so much to say on this topic, but I don't want to go on and on especially as I feel I would be preaching to the proverbial choir - I'll summarize:
IBM, please give us Configurable, not Customizable solutions.
Give Administrators the option to globally, or individually enfore application UI changes from Configuration, not Customization.
Give Developers the in-product tools neccisary to take their solutions from programs to applications - I should be able to control EVERY aspect of the UI when you're in MY application. I'm tired of being limited to "That's how Notes works".
Give users the ability to uniquely CONFIGURE their solutions. NTF doesn't use his To Do list, why should be be riddled with the UI share for functionality that he doesn't use.
The adoption and ramp-up time for a new hire would SOAR if they were given a "subset" of features and functionality that's available to us in the Lotus Notes client, instead of "here's the kitchen sink too - have fun!"
I honestly think the Engineers, Sales, and Marketing gang behind this product (and those of it's kind) take my advice: stop doing it just because you can!
Focus all of this energy into giving us something that we can work with, as the unofficial champions of your product, to provide our users and customers with solutions that WILL make them more efficient, more successful, and ultimately make your product a cornerstone in their day-to-day.
Posted by Chris Toohey At 01:30:12 PM On 04/20/2006 | - Website - |
One for example is not being able to do a correct summing in a view where you have "show response docs in a hierarchy" enabled. If you set this option you can not sum up a column with values from the main and the answer document. When asking IBM why this is so the answer was, because it slows down the building of the view. Hell yes if this slowing down the view and i want this to happen they will not tell you, "use this option in that way and you'll end up in losing 10 to 50% view performance" but they are simply restricting me to building either flat views and doing the sorting to keep related documents together by myself, or building an agent that fetches the values from the answer documents and puts them in the parent doc.
Another example is View headers, table headers and buttons. You can use formulas in actions, for computed text, outlines to do some tricks as we did with !!HELP!!. But this is NOT fully implemented and as it is not fully implemented it is in fact nearly worthless, because you cannot go down the complete road.
Speeking about multi language ability and views. Most of our notes applications make use of keywords. Most of us program keywords with aliases if they do something in your application. If you have users with different languages wouldn't it be nice to let them show this keywords in their language when they are displayed in views? And that this could be done without programming everything into the view? Yes this would be nice and no this cannot be done because IBM does not allow this (If anyone wants a proof for this i have a database to test this in several ways available). This is f....ing stupid.
Posted by Thomas Schulte At 04:56:00 AM On 04/21/2006 | - Website - |