Talk Page for TWikiUsability

I'm evaluating TWiki for my company (Main.TWikiStoryOfMaxFordhamLLP) and am currently focused on usability issues. I love the Admin side but am nervous about easy of use for new and occassional users. I'm writing this up at TWikiUsabilityTestingAtMaxFordham

-- TamsinTweddell - 29 Jul 2007

Twiki was my first taste of wiki and I have to say it works really well. The admin interface is superb.

One of the problems I have in selling it to those who are unfamiliar is that it has a reasonably unique markup and little by way of help for the day to day editors of the content. Having used it consistently for six months and an experienced coder I got used to it, but it was more awkward for those doing one entry, leaving it then coming back to it some time later.

As a "power user", twiki is very elegant. Based on some work by Barry Crabtree and Cefn Hoile, I have extended the stylesheet metaphor and defined the templates within the twiki as well. This allows the maintenance of the site to be done purely through the user interface.

However, for the basic editor, mediawiki with its graphical markup bar has a definite edge.

-- SamJWatkins - 09 Dec 2007

We are working hard to integrate a visual markup editor with the new release (4.2).

-- ArthurClemens - 09 Dec 2007

And if all you need is a graphical markup bar like MediaWiki - TWiki has several Plugins - the most current being NatEditContrib

-- SvenDowideit - 09 Dec 2007

TWikiUsability: Thanks to Carlo for yet another cool and useful article on usability, that gives a good overview. This article should be promoted on the Codev-frontpage.

-- MartinSeibert - 24 May 2008

I gave a TWiki-Heart for these valuable usability-contributions (TWikiUsability, UsabilityCategoriesAndHeuristics) to Carlo Schulz. He is our Usability-Champion! - TWikiHeart

-- MartinSeibert - 24 May 2008

Comments moved here from TWikiCommunitySummit2008Q3, regarding the agendum "A face-lift for TWikis user interface (improve usability)"

IMHO this is an absolute must have for a 5.0 release and it is as important as things like storage model and other technical things that happen in the background.

-- CarloSchulz - 14 Aug 2008

I've done a lot interface work for our TWiki at IBM which I'd like to demo. I'll try to transfer the ideas/designs into a more generic wire frame so that the resulting proposal wont depend on a specific skin. -- CarloSchulz - 14 Aug 2008

IMHO - we do not need a face lift and another attempt on destroying the wonderful user interface we have now. We should create new alternative skins. Not destroy the wonderful Pattern skin UI we have now. Users don't want the UI to change at each TWiki upgrade.

-- KennethLavrsen - 14 Aug 2008

Yeah, right. Ever thought about that there is a big difference between being perfectly used to an interface and an perfect interface? Don't try telling me the current interface is perfect - that would mean a 0 on your NerdoMeter. It's like me telling the techies in here we really don't need a new storage model, query language, meta data representation, caching or what ever. "Users don't want the UI to change at each TWiki upgrade." Who the hell is talking about changing it each time? 5.0 is a major release, not just an upgrade. If your argument was valid we would still have the Cairo interface. You are damn right about "We should create _new alternative skins"_. But have you ever asked yourself why there is a demand for alternative skins if the currrent one is already perfect? If you don't want an improved standard interface - fine, but you are not "the users". Ask Peter about the major concerns of his customers regarding TWiki - if you were right usability wasn't one of them. I'm sorry if my language is rude or offending. I fully respect your statement and personal opinion. I knew you would give such an answer but the way you did is quite frustrating. I'm offering help on a domain where I can help the TWiki Project the most. To me your statement reads as "thanks dude, but we don't need your contribution" simply because you don't need it. You reject it without even looking at a single proposal.

-- CarloSchulz - 15 Aug 2008

I'm with Carlo on this. The current user interface is quirky, slow, and difficult to use, and is most definitely an issue with my customers. The usability issues are not just related to Pattern skin, but also to the underlying architecture of TWiki. For example, the redirection to an "oops" page is an utter PITA for usability. So is oopsmore. I could go on.

Perhaps usability could be improved within the framework of Pattern Skin - maybe this will be possible, maybe not - but I for one doubt it.

Go Carlo! I'll be the first dev to sign up to your task team, even before I know what you want to do!

-- CrawfordCurrie - 15 Aug 2008

Carlo you were taking it for granted that we all want a face lift of TWiki without questioning IF we needed one. I react on that and I express my oppinion freely and openly no matter how many think the opposite!

• I'm not taking it for granted that each and everyone wants a face lift. I've just overheard so many times consultants talking about how their clients suffer from twiki's usability and interface shortcomings.

I don't want yet another change of the pattern skin so my users again has to get used to buttons moving around just to move them around.

• Do you really think that this is my unterstanding of improving an interface?

Why do so many XP users use the classic look (Win 2000 look)? Why do people not want Vista unless it can be made to look like Windows 2000? I just had dinner with two of my wifes business associates. Both talking about that they wanted to buy the XP downgrade from Dell so they could avoid changing user interface again again again. I - and many others - are sick over the UI changes done in IE7. I can not find anything. Tabs feature - fine! Moving the menus around. Not good. Once you are used to the positions of the user interface you hate it when it changes - just to change it. Is the face lift from IE6 to IE7 a good one? No. It sucks. Is the face lift from Windows 2000 to XP a great thing? Not for me. I like the simple naked look where my eyes can quickly find what I look for.

I just upgraded the browser on my mother-in-laws computer (IE6 to IE7). Now she is totally confused and cannot use it at all. Back to teaching once again. Did it give her anything new? No! Microsoft could have added the tabs and kept everything at its place. But no. They had to facelift the application and move everything around. The only really new thing - the tabs.

It is like some developers just have to change things because they can. Not because it is really needed.

I am not talking about oops pages (users do not even know about something called oops). I am not talking about deep down documentation or twiki apps documentation. There is plenty to improve there. Same with the dialogues in the Pattern skin and the more topic actions. I am plain and simple talking about

• Where is the edit button?
• Where is the print preview?
• Where do I see who last edited the topic?
• Where do I click to see the changes since last last revision of the topic?
• Where is the button for attaching a file?
• Where is the link to WebChanges?
• Where is the list of webs?
• 100% agreement on this one

As an application builder I am sick and tired of having to adjust the columnwidths of the tables just because the font keeps changing up and down in size. Lately I had to change table widths for schedules from 110 to 120 points just because the font was increased a tiny little bit.

As an upgrader I am sick and tired of having to redo my simple little tailorings of the WebTopBar.

What is wrong with having the edit button at the upper right? Why should it be removed or moved to the left menu - just as an example? Does this really improve the UI or does it just change it?

• Never said this is wrong.

Why not let the pattern skin be and create a new "next generation" skin for 5.0. The new installers do not have the history burden. The upgraders do.
Having a frozen Pattern skin and leaving it as a default extension that must work would also ensure that we do not AGAIN change the basics for the skins destroying the skins mods people have made and AGAIN making it impossible to upgrade TWiki.

• again, 100% agreement on this one

You are probably mainly dealing with new installations and do not offer suffer the upgrade pain to understand what it is like always having to spend days on getting things to work again just because some style sheet names changes and some styles changed.

But you all must have experienced the annoyance of getting a new version of MS Word to learn that the new version has moved everything around but cannot really do anything new for you.

Most of the facelifting is pure marketing. They need to change the look to make people think they get something new by upgrading. But in reality you just get annoyed.

• Do you think this is my intention? Marketing? For what? I don't earn any money with TWiki...

The last thing I fear is the constant suggestions to HIDE things. Those poor users get SO confused so let us hide things from them. Result. TWiki gets easier to use the first hour you use it. And a pain in the arse for the next 10000+ hours.

Why does a car have the buttons for fog lights, wind screen wipers, rear window heater etc easily available? These buttons should be hidden behind a panel in the glove compartment! That would make it much easier for the driver to understand the car the first hours of use. But a pain in the butt the rest of the time.

• Actually this metaphor is better than you might think. Remembering your very first driving lessons? I remember mine - horror. So many controls and buttons while I was already overloaded with just trying to drive. So what is the solution to this? Like you said it is definitely not removing these things. You need them once you mastered the basics of driving. But you really do not want to tell me that it is not a good idea to arrange the controls in a way so they don't get in the way of a learner? Making both parties happy is the goal of a good user experience!

Same with TWiki. TWiki is not a website people arrive at once per year to buy an airline ticket. You cannot apply the principles of this type of websites to TWiki.

• Ooh, really? And I thought I was using TWiki too...

For most of us TWiki is a daily work tool. Like in Word, Excel, Powerpoint etc we need the most used functions ONE mouse click or one key click away. Not hidden behind some menu. And they have to be visible so users learn that they are there. I see too much "smart" flash shit where you have to hover the mouse over areas and maybe you are lucky and discover some hidden feature. All looks so smart but works so badly when you are seeking information fast.

Like TV remote controls. The most daily use features have to be one button click away. And then the rarely used things can be in menus.

Please do not destroy my loving daily tool TWiki.

• I'm really wondering where your impression is coming from that I want to destroy things.

I think the pattern skin today has the right balance between what is visible and what is hidden in menus. It takes an hour to learn for the beginner and a couple of links at the bottom may not be obvious but not really disturbing either for the beginner. But after just one hour you love that the controls you need are quickly available at just one click away.

And just moving them around so they are placed different places is not going to play a significant role in improving the usability - is it?

But the re-direction to oops page. Yes that can be improved dramaticly. Same with many of the other functions. Example a good ajax app can make changing the parent much easier instead of scrolling 1000 topics to find the right one. We can improve this dramaticly. This is not what I worry about.

I worry about

• The number of and location of basic UI controls in normal view mode (usability of experienced users)
• The compatibility with tailoring of left bars and top bars (upgrade hell)
• The change of font sizes again so everything gets badly alligned in tables and similar. (application upgrade hell)

If we really want to improve user interface then it is when you have clicked Edit we should focus. You just have to pull a little in a table side with the mouse in Wysiwyg to get in trouble (it turns into HTML table and requires expert help to get it back in TML). It is when editing and building twiki apps. Changing the column width of a table is a pain for example (have to apply TABLE tag and know the syntax). This is where we should focus. Not move the edit button once again and change the font size and heading colour once again.

• Agian, almost 100% agreement

I hope you understand where I come from now.

-- KennethLavrsen - 15 Aug 2008

I agree with very much of what you say, Kenneth. I also get very frustrated when developers of any application move the entire UI around every new version (look at Acrobat for example). I do have one issue though that constantly bugs me and comes up almost every day with my users.

That is, there are many elements in the shipping PatternSkin that no users use or need. They are there exclusively for the benefit of the developers and/or admins.I'm talking about things like Statistics, Preferences, Web utilities, etc. I use these links every day. But I'm an admin. I know how to customize my LeftBar to add them there if I want to. The everyday user, even after more than a year of becoming familiar with TWiki, simply has no need for these items.

But they remain in prominent (read: valuable) positions, taking up space, and pushing other items that are used every day by every user to the bottom of the page or LeftBar. Why not remove these items and give the space to the everyday user? We admins and developers can add them back to our personal LeftBar.

I do agree that it's best not to go around moving things like Edit buttons.

As far as heading font sizes and colors, I would not propose a complete redesign. But the current PatternSkin defaults do not make any perceivable distinction between one level of head and another (except for the first and last). That's just not acceptable. Headings are critical to the organization of information. For them to perform that task effectively, they must be noticeably different--at least in size, and preferably in color as well. That's not to suggest that every head level should be a different color, but there should be distinction between one level and those most often used above and below it.

Here's an example of how this is handled in educational book design:

The key is that like information goes together. It is separated by more space between it and other dissimilar information, but by less space between the elements of the like information. There is some established visual pattern that defines the level of separation of the groups of information. Not an enormous difference. Not a rainbow of color or fonts. But just enough to tell, at a glance, how the information is organized. It's this that our users most find lacking. H2 through H4 or H5 are exactly or almost exactly the same size. The higher head levels use a roman font, but the lower head levels use a bold font. And the color of all heads is the same (this alone would not be a crime, it's just a more serious problem when there is no distinction in size and spacing).

-- DavidWolfe - 15 Aug 2008

Kenneth, see my comments in red. Actually we are not that far away from each other. IMHO it is mainly a communication issue.

A personal remark about your comment though.

Maybe I'm young but I work as a Consultant for Usability, User Experience, Information Design/Interaction or what ever you want to called it for some years already. It's my business and my pleasure too help improving things. That, and nothing else is my reason for being here on TDO. If you feel it's not needed or you don't want me too help - fine.

I never received any negative feedback for any of my work so far. I worked for small and big clients. I worked with different types of interfaces, web sites, mobile phones (motorola phones for example ), white ware. The list goes on.
There was not a single project that ended up with an interface that was worse than the one we had before the project. Our clients would tell us for sure if things would get worse instead of better.

I'm far away from being a guru in this domian, but trust me, I know my business quiet good. And I definitely don't need you telling me how interfaces work, what the results of changes might be, that different users have different needs. I'm not stupid.

And in this context I would like to ask you to stop talking to me like I just discovered the internet!

-- CarloSchulz - 15 Aug 2008

Kenneth, what's going on here? Did you have a bad time and feel like Carlo needs getting ranted all over? Please, pipe down and let the good guys do good things. Keep your Motorola users on PatternSkin, but let us move on.

-- MichaelDaum - 15 Aug 2008

I guess it is the word facelift that made me assumed yet another moving around the UI controls. That combined with some of the proposals I have seen lately for the twiki.org face lift where this is exactly what was proposed --- moving around everything.

If what Carlo wants is improve usability I am all in. But if it is moving UI around or hiding the controls like raw view, backlinks, history etc then I am against.

I guess we should use another word than facelift then because facelift triggers me to think the eternal moving around of controls we know and hate in Microsoft software.

-- KennethLavrsen - 15 Aug 2008

Kenneth, back off, please.

-- MichaelDaum - 16 Aug 2008

I totally support Carlo in this. The problem is that the functional interface has been frozen for years now because of the small develop windows we had (and in that time it was more important to fix bugs). Let's move it forward and create a better (new) user experience.

-- ArthurClemens - 16 Aug 2008

Micheal and Carlo - I do not know why my input is seen as ranting Carlo. I have explained that it is the word face lift I understood differently and it has been clarified that Carlo meant something else.

I hope back off does not mean shut up.

If we can all agree

• Make a new beautiful girl and not give the old one a face lift
• Ensure the old one still works and is part of the default distribution

then we all get what we want

Note in Windows XP how you with few clicks can choose a "classic start menu". If we can do this in TWiki many upgraders will be happy I am sure. All the user interfaces that happens AFTER you have clicked the first link can and maybe should still be common to all skins. I see no reason why changing parent, viewing the history etc should be totally different from skin to skin and if we want to replace the oops screens for sure we do not want to maintain this for pattern only. It is only the pattern skin in view mode I want to keep relatively intact. Everything else below can in my view be altered.

It should not be a total impossible task to keep just the view template for Pattern intact.

At both the first and 2nd TWiki summit I talked about freezing an API for skins including the essential CSS definitions.

We have seen that the pattern skin from 4.2 does not work in 4.1 and we have even seen Wysiwyg from 4.2 not working in a 4.1 because of small changes in the skin. When I joined the TWiki project Cairo had at least 5 skins that all worked and was maintained. Now there is 1 skin that works when a new TWiki is released (even classic is normally partly broken) and Nat skin gets repaired and working again months later as people report their problems.

It would be nice if people could contribute skins that would still work after next TWiki update. Like when you write a plugin, there should be an API you can stick to which ensures that your skin still works after next TWiki release. This does not mean freezing all skins and have no development. It just means that we document those skin related TWiki variables and a set of CSS names that a skin can depend on.

The 3 things we win from this is

• We will see more alternative skins in future because people will be more likely to make skins if they know they will work in future
• People will be able to make much more tailored skins that fit their corporate identity - and trust they still work when they upgrade TWiki.
• A pattern skin or Natskin that follows the API will also work in future.

I believe this is possible to do and still create a better new user experience.

-- KennethLavrsen - 16 Aug 2008

I would like to point out that MoveableTypeSkin still works on all twiki 4.x versions. While its not the prettiest skin at the moment, its look is entirely css based, so anyone designing a css look for it, would have their TWiki still look the same after upgrade.

It is already not impossible to make a skin that upgrades - I created MoveableTypeSkin to show the skin developers what is possible.

-- SvenDowideit - 16 Aug 2008

-- MichaelDaum - 17 Aug 2008

I am ordering that T-shirt. I love it.

-- KennethLavrsen - 17 Aug 2008

On its way

-- MichaelDaum - 17 Aug 2008

If you've read my comments (and maybe re-read yours) you should know what led to the impression of ranting.

As you've already proven a good sense of humor maybe this dilbert strips is helping you to understand it:

-- CarloSchulz - 18 Aug 2008

I just would like to comment, that TWiki definitely and urgently need an increase in usability, in case you want a broader range of companies and communities using it. Otherwise they move on to Xwiki or other sources. If you just want to have one of the best Wikis (if not the best ) and want to keep it a hidden secret then it is fine as it is.

-- WolfMarbach - 21 Aug 2008

I am all for improving usability.

It was originally a one liner on an agenda I reacted against because it said "give TWiki a face lift". And I understood face lift as moving some buttons around and changing some colours without actually making any improvements. We have later had clearly clarified that Carlo meant something else than what I understood.

There are lots of things usability related that can be much better in TWiki and I look forward to seeing and participating in making these changes.

I have stared working on a sort of requirement spec for what I think needs to be defined as an "API" for the skins and for the Pattern Skin in particular. I am about half way done and I will post it when I am have a full draft ready. Then you all will understand much better what it is I feel is worth protecting. It is not really that much I want to freeze. Stay tuned.

In the meanwhile try reading UpgradeTesting (read the comment from David Wolfe and Sean C Morgan). Interesting comments from upgraders.

-- KennethLavrsen - 21 Aug 2008

UpgradeTesting is definitely an interesting read (though it doesn't let you feel that good about TWiki ). Thanks for the tipp. However, the more I think about this whole UI thingy it becomes quiete obvious that we need a proper UserInterfaceStrategy. There are simply too much things to consider before we should even start thinking about a specific design.

-- CarloSchulz - 22 Aug 2008