Bad data bad decisions 2

Ages ago I moaned about how the wildly unrepresentative User Experience clicks program had been used to justify the abysmal ribbon. And have continued to ridicule it even though I said I wouldn’t – sorry about that.

Recently I had a pop about the wildly skewed on-line help usage data that MS are now using to try and slow (or maybe even reverse?) the decline in usefulness of Office Help. Assuming they continue to infer the behavior of the population based on their skewed on-line sample, then Help in Office 14 will likely be all but unusable for experienced users.

Imagine though for a minute you had accurate information about which file types people were working with. And further that many of the supported ones are almost never used. And also imagine some of those file parsers had some flaky pre-trustworthy zero day exploit fodder in them. Looks like a quick cheap security win right? make access to those files (and their flaky parsers) hard by default, and everyone’s a winner.

Meanwhile back in the real world, sadly the data is woefully skewed towards individual users with limited computing experience. Expert users and corporate users are barely represented – oops.

The User Experience piss poor data blight strikes again!

Does anyone have an example of where that ‘data’ ;-) was used and we got something useful?

Please Microsoft if there are any features in (or removed from – or just shuffled around in) Excel 14 driven primarily by those clicks, please get some feedback from the Excel community before baking them into the final build.

What do you think? Am I missing something?

cheers

Simon

David LeBlanc covers it very well here.

2 key takeaways from his blog post:

  1. The choice was leave everyone unpatched (/insecure) for months or do this quick and dirty now. I suspect if they had had more representative usage data things may have turned out differently. I don’t blame the security team, I blame the invalid inference from the massively skewed User Exp. ‘data’.
  2. The excellent way he deals with a rude commenter
Advertisements

11 Responses to “Bad data bad decisions 2”

  1. Stephane Rodriguez Says:

    Regarding user interface dumbed down for the masses, I think that you are way too kind about those who signed this off. I think this is just bogus design. Take a look at those two screenshots :

    That’s the dialog box of the axis formatting options in charts. Note that those options don’t appear in the Ribbon, so you have no choice but use that dialog.

    What do we see between the two,
    – the 2007 version is way more cluttered. And it even includes horizontal separators on the right side, illustrating an unfinished work.
    – everything renamed. This helps a lot…Not!
    – way more tabs.
    – note how the Line options are split in two tabs now, even though Fill options aren’t. (there are actual Fill color and Fill styles options however, it’s hard to understand the logic behind the split of just the Lines options).
    – the dialog shows up in the center of the screen, so unless you move the dialog, you probably won’t see that previews are applied immediately. (Note that previews are very inconsistent across the user interface. In many places, there is no preview at all).
    – there is a question mark button on the right-hand corner, which in theory is an improvement. But when you click on it, it’s not the typical Windows in-place help where you click on the control and then a floating tooltip explains what it’s supposed to do. What it does instead is automatically open another window which takes ages to open, and then the context is lost : the content of the help does not match the option names.

    It is hard to believe this was designed to be a productive user interface, neither a dumbed down user interface, nor a user interface for experts. So what is this user interface for? I still wonder…

    Regarding the security brouahh over file parsers, the next logical conclusion after disabling older file formats by default just in case that would cause troubles, is to block .xls files as well since the .xls file parser is known for being extremely fragile, susceptible to attacks based on cleverly corrupted files. Why do not .xls files go now if security really matters?
    Could it be because this is just nonsense?
    Could it be that Office 2007 files (.xlsx) are just as insecure as .xls files? (after all, aren’t binary blobs moved over there? Who is naive enough to think that just because there are XML angle brackets, it’s secure for the XML parser to read a file?)

    One thing that should always be kept in mind, something that Microsoft has been very silent about so far, is that Office 2007 alone contains a lot of new code, the migration code. Migrating .xls to .xlsx is new code, and there is a lot of potential holes there. Migrating .xlsx to .xls is new code, and again potentially a lot of troubles there as well. It goes and on an. Oh, by the way, Microsoft shipped separate components such as the compatibility pack : yet more new code. (apparently the migration with the compatibility pack is not a 1-to-1 feature match with the built-in migration of Office 2007, so the only conclusion is that these are indeed two separate code bases).

    If we assume security holes (and bugs) are directly proportional to the amount of new code, then the conclusion is quite different from what Microsoft wants us to think.

    Let’s also not forget that changes due to ISO are also going to add more migration code in Office 2009.

  2. Marcus Says:

    DBA’s grieving Excel file formats at JoS:

    http://discuss.joelonsoftware.com/default.asp?design.4.581096.7

  3. MikeC Says:

    You never know, we might see some change in the way the MS Office decisions are made in the future (ie Office 14+). As El Reg says in a marvellously-titled piece, Jeff Raikes, head of MS Office, is leaving this year, and we’ll have someone new at the helm (Stephen Elop from Juniper Networks – anyone know about him?), possibly with a different attitude to garnering opinion prior to release?

  4. Jon Peltier Says:

    MikeC –

    Wasn’t Steve Sinofski in charge of Office until soon before RTM, when he replaced Jim Allchin as top Windows guy? That makes two changes.

    As long as the Office tail wags the Excel dog, Excel users will wonder about design decisions.

  5. Stephane Rodriguez Says:

    Stephen Elop is Macromedia’s former CEO. When Microsoft makes a strategic investment in pushing Silverlight (i.e. Flash competitor), it’s not a stupid move.

    Funny nobody in tech sites mentioned it.They seem to just repost Microsoft PR shit and happy with it.

    As for jeff Raikes, I’m interested to know where will this guy land. I would hate to learn that he will just become another Microsoft proxy in whatever company he joins, just like this ex-Microsoft exec guy now working at BBC who, drum rolls, oversees the anti- customer Microsoft-DRM BBC initiative. I mean this guy is probably making BBC buy a lot of Microsoft software, contracts, … He’s an ex-Microsoft employee, and yet he’s turning a lot of money towards Microsoft.

  6. Stephane Rodriguez Says:

    Just a few things to add,

    Stephen Elop then moved to Juniper Networks (possibly after Adobe acquired Macromedia), and is now joining Microsoft.

    Steven Sinofski was one of the gazillion Office managers. I’m not surprised there are many persons, all overseeing Office, given how well designed it is.

  7. Simon Says:

    Jon – I totally agree.
    I reckon all 700 Office team members use outlook and word to a reasonable level, I bet only a handful use Excel (or Access) and probably less than half those use VBA. (and VSTO???)

    Stephane I think they blocked those files because their shitty ‘data’ suggested no one was using them. hopefully their ‘data’ is not so completely shite its telling them no one is using .xls. (next year maybe when they want everyone on .xlsx!)
    Are you hearing much noise about the upcoming ISO stuff? I was expecting to get drowned out in the commotion like when they were trying to get through ECMA, but it still seems quiet (apart from the SP3 excitement – not sure that will help their case?)
    Good point about silverlight, and the BBC debacle – although they overplayed their hand there and got a slap

    Did anyone else notice that servers were moving back out of Office?

  8. Simon Says:

    Jon – I totally agree.
    I reckon all 700 Office team members use outlook and word to a reasonable level, I bet only a handful use Excel (or Access) and probably less than half those use VBA. (and VSTO???)

    Stephane I think they blocked those files because their shitty ‘data’ suggested no one was using them. hopefully their ‘data’ is not so completely shite its telling them no one is using .xls. (next year maybe when they want everyone on .xlsx!)
    Are you hearing much noise about the upcoming ISO stuff? I was expecting to get drowned out in the commotion like when they were trying to get through ECMA, but it still seems quiet (apart from the SP3 excitement – not sure that will help their case?)
    Good point about silverlight, and the BBC debacle – although they overplayed their hand there and got a slap

    Did anyone else notice that servers were moving back out of Office?

  9. Stephane Rodriguez Says:

    “Stephane I think they blocked those files because their shitty ‘data’ suggested no one was using them.”

    I don’t think so. I think they have just refactored their stinking code a bit and thought it was about time to throw some of the old stuff. I expect them to set a trend and throw more file formats. But understand why they are doing that : no matter the changes they make in their own codebase, the public visibility we have of it is the branded products, and Office 2003 is now regarded by Microsoft as a product that is legacy (since Excel 2003 only produces .xls, and .xls is described as legacy, their words not mine). That’s part of their marketing initiative to get people scared of Office 2003 and therefore think of adopting Office 2007.

    “Are you hearing much noise about the upcoming ISO stuff?”

    Yep. I follow this very well (obviously, I’m interested in the outcome, if I have to spend another few months supporting even more variations of their XML spaghetti). If you have a question, don’t hesitate. What’s going to happen :
    – Feb 2008, final ISO meeting in Geneva : OOXML gets adopted. No question this is going to happen, Microsoft is bringing all their partners to the voting table. ISO is completely owned.
    – Problem though, OOXML (specs) is already at odds with the actual Office 2007 file formats (see my articles), and Microsoft Office bloggers have documented a number of changes they are making to satisfy national bodies and others. But those changes are even more at odds with what Office 2009 will do. I can already see what will be made available in Office 2009 from a file format standpoint just based on what they’ve done the last two decades. So for anyone supporting OOXML, he/she will have to support OOXML 1.0 (for compliance reasons), Office 2007, OOXML 1.1 (Feb 2008), Office 2009. That’s 4 variations just to get the mention “supports Microsoft Office XML”. Obviously, this will only be accomplished by some of the biggest vendors out there who can afford so much investment in it, the exact opposite of what they went ISO for.

  10. Harlan Grove Says:

    The affected 1-2-3, Quattro Pro and .DIF file formats HAVE BEEN DROPPED from Excel 2007. Microsoft isn’t going to spend ANY MORE TIME on the code that reads/writes files in those formats. The code still exists in Excel 2003, but I’d bet Microsoft is covering their collective butts making use of those file formats a ‘security issue’. If anything were to go wrong in the future, they can say ‘we told you so’ without having to make any effort to fix the parsers. And, surprise!, dropping DEFAULT support for these file formats could be perceived by many as one less reason not to upgrade to Excel 2007.

    As for ISO, one can hope the OOXML farce will undermine ISO’s credibility and lead to fundamental reforms, e.g., voting members would lose voting status if they fail to vote on at least 2/3 of standards ballots. That’d rule out most of the Microsoft shills that have become voting members recently.

    And to extend Jon’s pessimism about Excel’s particular future, now that the Excel developers have Excel Services to play with, Excel itself is likely to stagnate if not lose features while the UI grows. If Microsoft want’s data, they could use some text filtering on Excel newsgroups’ postings. They might see that people want such radical things as a function that could return a given cell reference’s worksheet name (without having to parse CELL(“Filename”,rng), and then only after saving the file), a function or functions that could be used for conditional counting/summing/etc based on formatting, a function that could return a number indicating which conditional formatting criteria may be met, functions for parsing text strings from right-to-left, 3D functionality at least in the sense that one could specify an output range for an advanced filter in a worksheet other than the one that was active when the filter command was initiated, and (very radical here) the ability to use a query result as the 2nd argument to lookup functions (something Lotus 1-2-3 provided more than a decade ago). Won’t happen (believe me, I’d love to be proven wrong).

  11. Stephane Rodriguez Says:

    I think the next big thing for Excel will be : 1) JITted spreadsheets 2) better integration with internet aware services (don’t think Microsoft will make it easier to integrate Google docs API, but someone else will…) 3) more reporting functionalities (business layer, etc.)

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: