OneNote Testing

A registry key for OneNote worth backing up

Looking through my Sent Items folder in Outlook for the past few weeks, I realized I had asked and told a few people about this registry key:

HKEY_CURRENT_USER\Software\Microsoft\Office\12.0\OneNote\OpenNotebooks

I've had to modify the contents of this key a few times for troubleshooting recently and I always keep this backed up on my dogfood machines. Let's take a look at this key from a black box point of view.

As far as registry keys go, this is as close to self explanatory as they get. This is the list of notebooks that OneNote will try to open when it starts. This seems reasonable and the notebook location is laid out as an easy to read string value. But looking at the sentence "This is the list of notebooks that OneNote will try to open when it starts" leads to some interesting test questions. An obvious one is "what happens if you try to open a notebook and OneNote cannot connect to it?" An example of this would be a server hosting a notebook being down or your computer not having network connectivity. The answer OneNote has is "it will use the cached copy of the notebook" so you can continue working in this offline case.

Let's say I wanted to test this registry key to see what OneNote would do if I make changes. I would start by removing notebooks from it - this seems like a reasonable first step.

Here's what I will do. First, backup my current set of keys from a test machine I use. This is what I have:

Now I will delete the #5 entry (that is a SharePoint based notebook on a server I have in my office) and delete my cache file (the .onecache file I have - this avoids "living in the cache" and any behavior associated with that. Since I'm black box exploratory testing here, I want to keep the variables to a minimum as I try to find out what will happen). Then I will restart OneNote.

I expect I will start with a blank UI ("fishbowl" in the OneNote lingo) and then I will sync the notebooks 1-4. And sure enough, that is what happens. I get the blank, gray UI and then the 4 notebooks draw. The notebook on the external file share (the one cryptically named "12") shows with the spinning green icon, and then sync completes. My cache file was built and I am ready to edit my notes.

The reason I have this key backed up is that on my dogfood machine I have 11 notebooks open. They are scattered across different SharePoint, DAV and UNC servers all over the place, and opening them after I wipe my machine and reinstall can be tedious if I have to browse to each server. This way I can simply import my registry file and start OneNote to sync them.

Further testing would focus on adding new notebooks, adding invalid notebook locations, deleting keys out of order, leaving the cache intact, etc… The test matrix would get pretty large, but starting OneNote is fairly quick, so working through it would not take long.

The end result of all of this is that if you want to start playing around with closing notebooks or otherwise using a big hammer (deleting the cache files) to fix problems, you probably can't go wrong with backing up this registry key.

Questions, comments, concerns and criticisms always welcome,

John

Mapping the OneNote team

A couple of years ago I hung a map of the world on a wall in the test team hallway. I simply like having maps around. Of course, hanging a map around here invites people to make more use of it than simply having a map on the wall, and someone eventually got some sticky notes and starting adding them to the map. The location chosen was "where I am from and folks would put a little note pointing to their hometown.  This includes just about everyone on the OneNote team – not just testers -  and here's what the map looks like now.

OK, it's a lousy photo, I know.

But just look at the tabs scattered around the map. The US and Canada are represented, of course, along with the Caribbean, South America, Africa, Europe, Asia and if you look very close, you can see a little yellow square at the bottom right pointing to Australia. If we can just get someone from Antarctica we would have a complete continental set.

Seriously, it is always great to have so many different viewpoints around. We get to see our product from the points of view of representatives from every part of the world and that keeps us from designs that only make sense (for instance) in one language or culture.

It also makes for great "hallway conversations" and fantastic pot luck dinners, too!

Questions, comments, concerns and criticisms always welcome,

John

I lost my Unfiled Notes in OneNote due to not planning ahead

One of the "tricks" I've learned over time for my dogfood machine is to set it up as if I know I will need to completely wipe the drive now and then. Wiping the drive is necessary when I need to change the operating system, for instance, or hit such a bad configuration state due to all the internal daily builds of software that I need to run that wiping the machine and starting over is actually less time consuming than trying to undo anything. For machines that only have one hard drive (which is all but one of my machines) I use fdisk to break the drive into 2 partitions and make one of them a C: drive and the other a D: drive. C: is for Windows and D: is where I put all my data, which includes a backup of a clean install of each version of Windows I may need to restore.

This worked well (for ten years) until about two weeks ago. I had to pave my machine and reinstall. I had all my documents and other data on my D: drive as usual, so I wiped the C: drive and started over. A day later, I had to get to a table I had been creating in OneNote which I knew was in my Unfiled Notes section. And that's where I got stuck. My Unfiled Notes had been saved on the C: drive in the default location (under My Documents\OneNote Notebooks). I had forgotten to change the location where all my notebooks were stored. Since all of my other notebooks are shared and/or stored on different network locations, I wasn't worried about them, but losing that one table was tough.

Here's what I'll do to avoid that in the future. Click Tools | Options and then the Save tab along the left.

I can modify where my Unfiled Notes get stored from here. This is a bit trickier than just pointing to a new folder, though, since the target is the actual "Unfiled Notes.one" file. I'll move that file from its current location using Explorer to where I want it (in my D:\images folder, which is where I keep my backed up copies of Windows - I know I'll never delete that folder) and then point this location to that moved file.

So now when I have to wipe my dogfood machine and restore Windows, my Unfiled Notes file will be intact on a separate drive. To point my next installation of OneNote to that location is easiest done by exporting the registry key this setting creates and re-importing it (rather than going through the UI). That key is at

[HKEY_CURRENT_USER\Software\Microsoft\Office\12.0\OneNote\Options\Paths]

"UnfiledNotesSection"=hex(2):<<<bunch of hex numbers that don't make much sense out of context>>>

The path can't be just a string (as opposed to String, for you .NET types out there) since it can contain Unicode characters, so it has to be stored in a somewhat unusual format.

Next time I need to restore, my Unfiled Notes will not only be intact, OneNote will get pointed to the previous file automatically.

This is kind of an unusual workflow, admittedly. I would not expect many people to need to reformat their hard drive often so this may not be useful to many folks. But it does expose some of the functionality within OneNote and lets me test a way to ensure I don't lose my data for this extreme scenario.

Questions, comments, concerns and criticisms always welcome,

John

Using Windows accessibility features to change the OneNote UI

Last week I showed how you change some global Windows font and color selections to make changes to the way OneNote displays. I also explained why you probably don't want to do this. Then over the weekend, I saw this question on the discussion group:

"How do I change the background colour of a page to black?"

A short answer that I gave is to change Windows to use black as the window color. This means that most applications (unless they handle their own colors) would start to use black. Notepad is a great example. While this is by itself not bad, unless you remember to change the font color, you will wind up with black on black. This is clearly not ideal.

Instead, you can take advantage of a feature we have tested - Windows Accessibility. Depending on which version of Windows you have, you can change to one of several different Accessible modes that use high contrast colors to make reading the screen easier. For instance, you can change to High Contrast (Black) and this is what OneNote will look like:

You get the black page color, but a whole lot more functionality as well. The page title tab gets changed to gray, the font becomes white and even the selected text gets highlighted in purple. This is a much better choice compared to finding and tweaking all these colors yourself. Notice all the powertoy icons do not have their background color set to transparent? This is not a OneNote bug, but looks like the powertoys could use some updating if the UI for them is critical. The powertoys that use text labels instead of icons work as expected, though. Yay for plain text!

The default high contrast theme on Windows XP is very similar:

You get green highlights and the you can change is yellow.

The white theme is not quite the "reverse" of the black:

The section name here shows as a link, but otherwise the change is fairly obvious.

So what I'm going to do, since I brought these modes up, is to use the high contrast black theme on my tablet. I like the dark background, and since the ink color is set to white in this mode by default, it's like writing on a chalkboard.

I'll also track down the bug in the UI shown above. Can you find it?

Questions, comments, concerns and criticisms always welcome,

John

You probably don’t want to do this to OneNote, but here’s how

I saw this question on the Outlook Newsgroup a few days ago:

"Is there anyway to change the font size in the main outlook screen ( inbox...)? [I] want to keep the system font size..."

And one of the replies said there was no way to change the font in the folder list of Outlook. I had looked at this a while back from a OneNote question which was similar (the person wanted the change the section name font along the top of the page) and I hesitated about explaining how this is possible since it can lead to odd behavior across all applications in Windows but I finally decided to point out how to make changes like this. I want to show you how you can change these font settings the application doesn't expose. Then I can show why I don't recommend doing this.

Windows has settings which let you control which font is used for the UI. You can see which fonts Windows is set to use by opening the display properties | Advanced dialog. This is what it looks like on Windows XP:

On this particular system, the Menus in Windows applications use the Tahoma font at an 8 point size. It just so happens that this is the font (the Menu font, that is) used by Outlook for the folder list. I found this out by “black box testing” also know as trial and error.  If you want, you can change it to use any installed font, but be careful. I changed it to 24 point and this is now what Outlook looks like:

So the font in the folder list changed along with the menus. It also looks like the command bars use the Menu font. The Search address books control uses a separate font, and the font in the view is controlled by Outlook view settings (View \ Arrange By \ Customize Current View gets you to those font settings). To my eyes, this doesn't scale well and since it will change the menus in all applications on the system (well, those that let Windows handle the fonts, at least), I'd need to go through all applications to see if they are still usable.

One thing I did notice is that the Office theming overrides the font color so I always get black fonts with the blue color scheme. That makes sense to me - if I change the font to a color that I'm using in Office as a background, the font would be invisible. The little previewer Windows shows helps prevent that, but since the background in Office is set via a theme instead of the Windows setting, it becomes incumbent on us to help make the font visible.

Then for fun, I installed a 1970s era computer font, changed it to red and 24 point and looked at it with a PDF viewer (Foxit):

They use the color I specified but notice the large Close, Minimize and Maximize buttons in their command bar. If you look very closely at them, you can even see the background color for those buttons is not the same as the background of the command bar any longer. And if I really wanted to get picky (which I do want to do, since I am a tester), I could point out the icon to the left of the File command gets pixilated when I change the system menu font to 24 point.

So that's why I don't recommend changing these settings. What may look good in one application can cause odd problems in another. If you really want the large fonts, try the accessibility settings.

Last item: here’s what OneNote looks like (in case you want to change the font the section tabs use). I used the computerfont and changed it to 24 point:

So our search control uses the system font as well as our command bars. This makes addins that use text labels instead of icons a little easier to create (one less thing to worry about). I actually tried using OneNote with this setting for a day or two but quickly gave up. Segoe and Calibri fonts are just easier on my eyes, and that's what I use.

The computer font is available at http://www.dafont.com/computerfont.font

Questions, comments, concerns and criticisms always welcome,

John

Did OneNote cause this? (Haiku included)

This is not so much about testing as life at Microsoft. Near our coffee machines is a little bench where you can sit and talk. And at the end of the bench cushions is a little area that serves as a coffee table. And about three months ago, this appeared on the coffee table part of the bench:

A little abandoned notebook. I figured when I first saw it that someone had simply left it there and would return to pick it up, but no one did. I opened it to see if there was a name, but there wasn't. Only a few doodles and notes. As the days turned into weeks, then months, it has remained there, unclaimed.

I pointed it out to one of my fellow OneNote testers last week and he wondered if the person had dropped it there specifically because of a switch to OneNote. Maybe this person had made the transition away from paper and will never need this notebook again. I wonder.

A haiku to the notebook:

Abandoned notebook,
Are you to be remembered?
Did OneNote oust you?

Questions, comments, concerns and criticisms always welcome,

John

Looking at design types of “bugs” is a challenge

One of the types of "bugs" I tend to report now and then tend to stretch the definition of "bug" a bit. At first glance, you could say bugs are code defects - any time the code does not perform to the design specification, you are looking at a bug. Fair enough. But there is also the case that testers need to report when the design itself needs to be changed. Sometimes what looked good in the design stage may need to be changed once you see the results on screen. It is also possible to develop a poor testing habit when seeing these types of problems. Lets turn back the clock to look at a design choice from the past.

Windows 3.1 had the idea of color schemes. One of the color schemes was "Hotdog Stand" which was a very (to my eyes) garish blend of red and yellow.

We could debate the need for this scheme. I could imagine it was used as an example scheme of what "not to do" with customizing the scheme, though. If you dig into the dialogs, you can customize each of the colors windows displays and perhaps this was used when Windows 3.1 was being shown about what can happen if you make color changes just because you can. Lets go ahead and say this is implemented exactly the way the specification said to implement it.

I could also imagine that no one ever intended this to be used for any "serious" purpose. After all, the name of the scheme is "Hotdog Stand" and not "Law Office Interior."

The last bit of information about this scheme is that it was not the default. If I remember correctly, you had to open about two windows and two dialogs to get to this selection.

So we have a situation in which the code is doing exactly what the specification calls for. Bear with me, because we are now talking about behavior that is closer to 20 years old than 5, and history has already decided if this was a bug. So to make my point, I want to propose this scenario. The design team decided Windows 3.1 should be a "fun" release and want to make this the default that every one sees, and it did seem fun and whimsical.

Then you actually get to see it in action. You go with the concept for a day or two, then you turn it off and install a more moderate theme. Since you are on the team, this is a relatively trivial task. This is where you start to develop the bad habit - you are actively working around a design feature of your application. As the days go by, you can develop the habit of changing to the scheme with a utility, or muscle memory takes over and the first thing you do after logging is execute the keystrokes to turn this off.

At this point, you can't say there is a bug. You may be the only person who does not like this decision, but there needs to be some default. It could be the case that you simply don't like this design.  Again, the feature works perfectly, causes no problems and does exactly what it is expected to do.  The only “problem” that exists is the color selection. 

You may look around one day and notice others are doing the same thing. If you let the habit of working around the bug remain in place, it could cause you to not see the problem (I created) here: no one is using the feature.  But if you do notice the situation in context, it gets interesting and it’s time to gather some actual data.  You can't rely on anecdotal stories about what people are doing. User studies can be useful, beta feedback and other surveys can help find what percentage of users do not like the design. Then you can file a report (likely a bug) to reconsider the design.

Testers are on the hook for many different deliverables and too many times we can get wrapped up in "code" bugs and overlook the design implications. It's a balancing act that we must complete in order to ship software.

Questions, comments, concerns and criticisms always welcome,

John

How you can give feedback or report a bug with OneNote

Every so often, I get an email or report on my blog about a suggestion, criticism or bug report about OneNote. This is great - I look at a criticism as a chance for us to do better!. I figured I would show another way you can give feedback or bug reports in case you don't want to contact me directly. I'll also point out an advantage of using this method as opposed to just emailing me (or anyone on the product team).

We have a website at http://connect.microsoft.com that lets anyone make reports. You need to sign in with your Windows Live ID account (I won't walk you through the signup process - it is pretty straightforward) and then the top of the page will show this menu:

Click the Connection Directory. And while I'm looking at this, the UI is subject to change, so if it changes, let me know and I can update this image.

After you open the Connection Directory, on the left you can see a list of categories. Since OneNote is part of Office, click the Office link:

Almost there. Near the bottom should be a link to the OneNote Connect site.

Click Apply Now for whichever version of OneNote you have, and you get the OneNote page:

(Notice that blog listed last :) )

To report a bug, just click the Submit Feedback link. You will get a page that asks for basic information, and I want to focus on the last few questions:

Reproduction Steps are the steps I would need to take to reproduce the bug on my machine. Generally, we call them "repro steps" internally, and we make the exact same report when we file bugs internally. We also try to find the minimal set of steps to surface the bug - our goal as testers is to isolate the bug as much as possible to make troubleshooting easier.

The "expected results" is self explanatory. It is the behavior you expect to see. What is missing here is the "actual results" you see at the end of your repro steps. This can be anything from a crash (at one extreme, ie, "I click the button and I crash" ) to a missing punctuation symbol (ie, "The last sentence of the dialog does not end with a period."). You can also upload files. When you complete the bug report, it gets copied to the database and then I can see it!

Another question I see over at the OneNote Discussion Group is asking how often we look at the reports. While I can't say it is daily, we do keep our eye on it constantly. As one example, when we were trying to track down the onenotem.exe process going away, we scoured through there looking for commonalities. Our design team also looks through it to find areas that many people are requesting, and if you submit suggestions, others can vote on them. We try to incorporate those requests into new features. Finally, we (the test team) will dig here looking for bugs that we can reproduce, and we will port them to a separate database. These are the advantages that contacting me directly does NOT have.

Anyway, check out the Connect site. It’s one more way you can help OneNote.

Questions, comments, concerns and criticisms always welcome,

John

Looking at a bug we fixed in OneNote 2007 SP2

We changed the way the OneNote system tray process works a little bit in SP2. I want to explain the behavior we were seeing and show the bug fix we took.

The system tray process is the application that handles screen clipping and side notes. You can see it visually by looking down in your taskbar and seeing the little OneNote icon:

In task manager you can see this running as "OneNoteM.exe" (notice the "M" at the end). It doesn't take much memory, either:

When you create a screen clipping, this application handles the WINDOWS+S key (in high level terms) to invoke the actual screen clipping command.

Now with a little background, we can understand the report we get now and then that "I tried to make a screen clipping and it did not work." Without some context, there is not much we can do about this report other than assuming that onenotem.exe was not running. Since we have a little background information, we can ask some questions. And one thing we did indeed notice is that in all cases the system tray application was not running. This explains why the screen clipping does not work, but it only really only leads to the next question: "Why is onenotem.exe not running?"

There are a few explanations for this. To start tracing this problem, you can check your startup folder. Here's what mine looks like:

You can see the screen clipper and Launcher there, and looking at its properties does show it points to OneNoteM.exe. This is the command line: "C:\Program Files\Microsoft Office\Office12\ONENOTEM.EXE" /tsr (the /tsr means "terminate and stay resident."  Someone mentioned that acronym may be too old for most folks to remember.).

So one explanation for why screen clipping may not work is that the shortcut to start the application got removed from the startup group.

Obviously, if the application crashes, it would also quick working. And since it gets launched only when the machine starts - unless you start it manually - it would not be running.

A third reason would be that the user removed it completely, and this is the behavior we changed. If you right click the icon, you get this menu:

You can see the option to remove the OneNote icon. If you select this with Office 2007 SP1 or newer, we would remove the icon from the system tray and set the onenotem.exe process to not automatically run any more when Windows is restarted by setting a registry key. This behavior does not match 100% with the text - the text implies that only the icon is removed, and says nothing about whether the running process should remain around or not when you restart. It is possible to make that selection and think "All I'm doing is cleaning up my taskbar and killing this one process temporarily" and not realize you are changing the way this application starts. This is not the way other applications like Instant Messenger or Communicator and such work - they only get shut down temporarily and then restart when the OS restarts.

So the fix we took is to leave the setting to start the process startup routine intact by not writing the registry key. In other words, removing that icon will remove the icon from the taskbar but leave the setting to start it when Windows starts. The next time you restart, the screen clipper functionality is restored.

We expect this to help with the third reason I mentioned above about why the application is sometimes gone. There is not much we can do about number one (you may have specifically removed it for optimization purposes, for instance), but there is work we can do about the second reason. I will talk through "Watson reporting" and show the information you can give to help us next. The short teaser is that when an application crashes you get prompted to upload the crash data to Microsoft and we use that to analyze potential fixes for our products.

Questions, comments, concerns and criticisms always welcome,

John

Using OneNote as an email creator lesson learned

If you've been reading this for awhile, you know I spent some time using OneNote to compose all my emails. I received an email from a tester on my team last week and needed to forward it. She had used OneNote to compose the original mail as well, but I did not notice that to being with. Instead, the first time I noticed it was when I wanted to forward her mail and needed to re-format the first line of it. I could not format it the way I wanted and thought I was hitting a bug that needed to be reported. I had never hit this myself. But it turns out it wasn't a bug so I figured I would explain what happened.

First, here' s what I wanted to do. I received her mail and noticed there was a line of text "way off to the right" that I wanted to move back to the left. Here's what I saw when I wanted pressed forward and clicked in the body to the left of the line I wanted to move:

When I put my cursor to the left of the word "to" I wanted to press the "Decrease Indent" ribbon toolbar button to move the line to the left. Pressing it did nothing. I had never gotten behavior before and was confused.

Look at the top of the email inspector frame: the caption says "Table Tools." That was a clue. Tables are a method of forcing a layout scheme onto a OneNote page (we're typically pretty loose with letting you type or insert content any where you want, but sometimes you need information right there). I didn't know where the table was coming from since all the mails I sent did not have it.

Since she had sent me an HTML formatted message (this is the default in Outlook 2007), I was able to right click the original message and view the HTML source:

Looking at the HTML shows a table holds the line I wanted to move (here's an edited copy of the HTML table):

The last clue here is that Outlook uses Word as its email editor. So what I was seeing was the expected behavior of changing the indent of text in a cell of a table. When I moved the text to the left, it moved to the left most edge of the cell. Since the table was invisible, it did not immediately "click" and I had to look twice to see what was going on "under the hood."

The reason I had never seen this with my own mail is that I had not inserted an outline to the right of the date/time and title area. I had been keeping all my notes under that area and therefore never needed the table to be created to ensure the text was in the correct location when it shows in email.

Questions, comments, concerns and criticisms always welcome,

John

Some OneNote follow up items while my tablet is gone

So my tablet PC is still being used for debugging and it is taking longer than anyone expected. No big deal, except that my notes I had collected about setting the page size are on that machine's hard drive. I didn't have time to copy them anywhere for backup before I lost the machine. I'll spoil the ending by saying I had trouble setting the height attribute to zero if there was a page title. The easy workaround was to set the height to 18 pixels if there was a page title, which you can read from the XML.

So a quick note on other follow ups and some miscellaneous information:

1. I am waiting for the logging I added in April to help track the failing automation script. This was a very intermittent error so I am not surprised it has not been hit since then.
2. Ironically, I checked in some other changes to our napkin math automation just today.
3. While my Lenovo is being debugged (and the problems it is hitting at this point seem to be unique to that piece of hardware) I have a loaner Fujitsu Lifebook tablet. It's roughly equivalent to my Lenovo, but has a 64 bit dual core CPU instead of the 32 bit I had before. Currently I have it set up with Vista 64 bit and hope to get my other machine back before I need to change configurations.
4. I'm still puzzled by the Task Request from Meeting Notes powertoy being so frequently downloaded. I've heard the requests from our discussion group, the Connect site and direct email that we need more task integration, so maybe that is feeding these downloads.  I never expected it to be as popular as it has been, though.
5. Just about 100% of the program managers (designers of OneNote) that walk by my office regularly slow down to read your tweets. Many other folks from other teams do this, too. It starts quite a few conversations about tweeting, what the red/white/color scheme means, and what information you can get from it. (And that last topic there has a whole world of analysis behind it).
6. And some news that is almost marketing hype (but this Technical Fellow is very much worth listening to):

<hype>

On Thursday, June 18, Mark Russinovich will host a virtual roundtable for IT pros worldwide to discuss Windows 7 Application Compatibility. This is a chance for IT pros to learn all about the Windows 7 Application Compatibility and answer any deployment and management questions that they may have. 

Save the date and join Mark on Thursday, June 18 at:  https://ms.istreamplanet.com/springboard.

As part of the “virtual” experience, you may submit your questions about Windows 7 Beta to the panel live during the event—or submit questions in advance to vrtable@microsoft.com.

(I'll even add an ical attachment - 8K in size - below my signature on this blog if you want to add the date to your calendar).

</hype>

I hope to be able to broadcast the other information I have from investigating the height and width for pages from our API when I get my hard drive back. Gathering the information took quite a while and I would hate to go through that process again - I'd rather move onward!

Questions, comments, concerns and criticisms always welcome,

John

Looking at the XML schema for page width and height in OneNote

I received an email a short time ago which had a request for a particular page size format. In short, the college student wanted to have page which was a fixed width, but infinite height in order to take all notes for a semester long class on one page. Sadly, that is not possible, but I thought it may be interesting to see what I did to confirm this. In the process, I can demonstrate a workaround that may be "good enough."

First, I started the ONOMSpy tool. This is what I use to quickly modify the XML for a given page in OneNote. I also ensured I had the XML schema reference downloaded and was ready to go.

First, I opened the schema reference to look at the OneNote XSD. I looked up the PageSize schema:

I can see from here that the PageSize XML node exposes the Dimensions element, so I am in business.  (And sorry for the image – the XML formatting gets lost when pasting into Live Writer.  I will leave a bug on http://connect.microsoft.com.  In the meantime, OneNote screen clipping to the rescue!)

Looking in the Dimensions element of the schema shows two pieces of information. First is the actual schema:

<Dimensions
        height="PositiveDecimal"
        width="PositiveDecimal"
/>

Which shows the height and width of a page are individually exposed. This made me slightly optimistic that I could set a fixed width and infinite height. Then I noticed the limits on the parameters:

And here's where the request became impossible. The minimum value of the height is zero, but it has a maximum height, albeit of one million.

And here's where "good enough" may come into play. I can set the width to whatever width the user wants (let's say 1000 for sake of argument) and set the height to 1000000. This may be satisfactory if the user never lets his pages go over one million pixels in height. If the real constraint is that he wants all of his notes for a given period (one college class) on a single page, this may well be enough since the data set he's working with is finite.

I'll go through some testing I did for this next. Stay tuned.

Questions, comments, concerns and criticisms always welcome,

John

My tablet with OneNote is gone! (But here’s why it’s OK)

When I got to work today this is what I saw:

A missing tablet from my docking station. I was actually happy to see it missing (it had not been stolen, nor had more hardware gone bad), and here is why.

A short while ago, I installed a particular build (version) of Office. All appeared well, but after a few days of use ("bake time", in the lingo) I noticed some rather severe performance problems. Random hangs for 2-12 seconds, typing was sluggish, I couldn’t switch between Office apps and things like that. Other applications like Notepad or Wordpad worked fine, so I was pretty sure I was looking at an Office bug rather than a hardware problem. I went through the standard troubleshooting steps of getting the smallest set of steps needed to reproduce the bug, got a memory dump and a some traces of the soft hangs and filed a bug.

Quick note: a "soft hang" is one in which the application appears to freeze for some amount of time, but eventually the user gets control back. In my case, the soft hangs would last only a few seconds, but I was getting them so frequently that I could not really use the machine. A "hard hang" is one in which the application never returns control.

Anecdotally, I asked around to see if anyone else was using my same configuration and found one other person on my team who was, and he was hitting the same problems. Great - more data to troubleshoot.

After a few days, I checked to see what was happening with the bug. As it turns out, I was the only person reporting this problem out of hundreds of users with the same configuration. That itself is a clue - it may be unique to the personal tweaks I apply, something to do with the exact hardware configuration I use after all (I use ink on my tablet, for instance) or there may have been an error when Office was installed that was not reported correctly. At this point, I have no idea what could be causing the performance problems, and the developers from the setup team came by yesterday to investigate. They could not make any headway in the short period of time they had, so I volunteered my tablet to them to take to investigate.

I know all of this sounds a little vague and is missing some details. At this point, though, I don't have any details of the problem. I can only point to a single symptom and don't want to speculate on the root cause.

Taking hardware like this is not unknown, but it is not common either. Since we had such a small set of machines (2) that can reproduce the problem, it was the quickest way to get the needed information to the team to investigate the problem.

For now, I'm back to pen and paper for meetings. Sigh. I miss OneNote already…

Questions, comments, concerns and criticisms always welcome,

John

The Scientific Method and testing OneNote performance

We finished one of our performance rounds of testing a few weeks ago (one of many) and were looking at our numbers.  In some cases, the numbers looked very good and others showed some "areas for improvement."  One of the folks involved stopped by and mentioned that since all the numbers weren't good enough, our round of performance testing did not meet our goals.

I completely disagreed.  The goal of the performance work was to establish an accurate measure of OneNote performance, and we managed to do that.  When we set out to gather the numbers, we followed the Scientific Method of testing the system:

1. We asked a question:  how many bytes does OneNote send over the wire to the server when syncing a notebook?
2. We did our research.  We looked at the number of bytes of the wire sent by previous versions of OneNote and used them as a baseline.
3. We constructed a hypothesis.  We assumed that the number of bytes over the wire would be roughly equal to the size of the notebook in bytes + the size of the headers sent in bytes + the size of the authorization calls in bytes. 
4. We used the fiddler tool (www.fiddler2.com) to analyze the client requests to our SharePoint server.  This tested our hypothesis.
5. We analyzed the data gathered and drew conclusions.  Our conclusion was we could do better - you can't have too much performance!   
6. I sent a report to our stakeholders with our summary.

This looks like a successful trial to me.  We started without any data, followed a plan and quantified data that we know is good. We can act on the data and make changes to OneNote to try to get desired results.  Now that we have our tools created, we can verify the results of changes to OneNote as time goes by and measure performance.  We can use these measurements to make further decisions about OneNote.

The mistake the fellow made was assuming that since we did not have "perfect performance" from the start, we had failed in our testing.  That misses the point of testing.  We need to be able to gather meaningful, accurate statistics that tell us the stability of the product.  Failing to get meaningful numbers (in this case, perhaps using a known faulty stopwatch to measure time) would indeed be a testing failure.  My co-worker misunderstood the intent of the testing performance focus week.  We wanted to ensure we had a solid baseline from which to analyze changes in the future.  Now we do, and that is success.

I guess what I'm trying to get at is as a tester, I'm interested in knowing the tests I run are telling me about the state of the product I'm testing.  I'm concerned about the product, of course, but my primary role is accurately reporting that product's current state.  If my tests aren't telling me anything about the product, don't deliver consistent results or otherwise give me inaccurate data, then I am very worried and need to rethink my tests.  I would not be able to complete step 5 above - I could not draw any conclusions.

Likewise, I want to steadily progress through my testing.  It's pretty easy to start "testing in a circle," skipping step 2 of doing research beforehand and never verifying step 4 by never realizing you never had a hypothesis to test.  This probably deserves some more thought.  I'll try to follow up in the future.

Questions, comments, concerns and criticisms always welcome,

John

The downside of living in the cache

Way back when my server was up, then down, then back and such I was "living in the cache." This means that OneNote was letting me edit my notebooks while in "offline" mode and would keep track of my changes while the server was gone. I wrote a little about what this means and how we test it back in January. This gives me great flexibility to keep working and upload my changes when the server comes back.

But that configuration is not without its own risks. Had my hard drive with my cache completely died, or my cache gotten deleted or otherwise corrupted, I would have lost some of my work. This could also happen if I choose to delete the cache. I figured I would publish the list and add the "safety net" OneNote employs to help with this.

First, the list:

1. Outbound changes
• This means any of my changes that I had typed and waiting to get sent to the notebook on the server.
2. Misplaced Sections & Real Time Collaboration Sections
3. Notebook renamed names
   • If I had renamed a notebook so it shows a different name in the navigation bar, that new name would be gone. I actually do this quite a bit to shorten the names to fit more in the UI. For instance, our "Test Team Notebook" becomes "ttn" in my navigation bar.
4. Order of which notebooks appear on the notebook navigation bar
5. Your state of where you were in a notebook/section
   • This is the page you go to automatically when you click a notebook. It should be the most recent page you were on - you just go back to where you last were.

Items 3-5 are certainly nice to have, but I'm willing to live with the risk of losing that level of detail in the event of a hardware failure or crash corruption on my end. It would take only a few seconds for me to reconstruct the information if it were to disappear.
1 and 2 are more important. Looking at 1, I lose any information that exists only in the cache (changes I haven't been able to send to the server since the server is not there while I'm offline). If I'm not comfortable with this risk, the "safety net" I mentioned earlier can help me. OneNote makes backups for me automatically to files outside of the cache. You can see them in File \ Open Backup. This backup gets run once per day by default, but you can control that by clicking Tools | Options | Backup:

If I'm really worried, I can set the backup to happen every minute. And I can obviously click the button to force a backup whenever I want. This operation will take all the changes that exist only in OneNote's cache and apply them to the .ONE files on my machine.
Fortunately for me, despite losing three hard drives at once, my drive that holds my cache (and the cache itself) have survived unscathed.

Questions, comments, concerns and criticisms always welcome,
John