Note: I work for Microsoft, in the Office division, but I don’t work in or with the Outlook team. I don’t have any specific knowledge about their decisions or plans, and this post is based only my own experience here at Microsoft.

The web community has been up in arms the last day or so about a campaign via Twitter pushing for Outlook 2010 to stop using the Word rendering engine for HTML email. I engaged in a short friendly discussion with my buddy Vlad on Twitter about the issue, and that got me thinking about the issue a bit more deeply. And the more I thought about it, and read everything that was being written, the more I realized that people aren’t actually communicating effectively, and that pisses me off…

The fixoutlook.org website says this: “Microsoft has confirmed they plan on using the Word rendering engine to display HTML emails in Outlook 2010.” (emphasis added) Based on that, and the rest of the text on the site, it seems like the big beef they have is that Word’s rendering engine for HTML is not up to snuff. Fair point – it isn’t. But they contradict themselves in their updated response to Microsoft’s response when they say, “We’re asking that the HTML produced by the Word engine be standards compliant. This in turn will ensure that the engine will correctly render standards-based emails.” (emphasis added)

Wait a second. Do they want an editor that produces HTML, or a rendering engine that works properly? Doing the work to make sure the editor is producing clean markup might produce a rendering engine that renders nicely, perhaps it even should produce one, but that doesn’t mean it will. Those two pieces of work are not the same thing.

Lots of people have been pointing to a post from Zeldman about this. Zeldman’s a brilliant guy, but he completely misses the point when he perpetuates a rumor about why Outlook chose Word’s engine: “Rumor has it that Microsoft chose the Word rendering engine because its Outlook division “couldn’t afford” to pay its browser division for IE8. And by “couldn’t afford” I don’t mean Microsoft has no money; I mean someone at this fabulously wealthy corporation must have neglected to budget for an internal cost.”

Ummmm… no. In my experience, the phrase “couldn’t afford” at Microsoft R&D has nothing to do with monetary cost. It has to do with engineering cost. Features don’t exist just because you want them to exist. (An aside: Raymond Chen has a great quote about this that I can’t find. If you have a link, please let me know.) So I imagine that the rationale in the Outlook team went something more like this:

Well, we want Outlook to produce really rich emails, and we want it to have a really familiar look and feel, so people already using our products will feel at home. Hmmm, building an editor is extremely hard to get right, plus we, being part of the Office suite, have several editing tools already. Emails tend to be a lot like documents, so Word seems to be a reasonable choice. This way we can leverage all the editor features Word already has, plus things they’re building this release, and focus on the Outlook-specific work we have to do. We don’t want to invest our engineering time in building an editor when we already have one.

From this standpoint, it’s easy to see why using Word for rendering as well is the next logical step. You could argue that they should use IE (or Gecko, or Webkit, as Zeldman does, but again, just because those engines are free doesn’t mean using them is without cost) to render email, and Word to author it. That’s a reasonable idea. In fact, the Outlook team seems to agree with you, because if you get an HTML email that looks wrong, you see a link to open it in a browser. In fact, even the example image at fixoutlook.org has the “info bar” at the top that does this.

You can take issue with the implementation, certainly, because it sucks mightily. The fact that I have to open the email in a browser separately blows. I shouldn’t have to do that. But don’t kid yourself into thinking that integrating an IE window into the message window is so stupidly simple that Microsoft is maliciously avoiding doing it to somehow screw users. It might be straightforward – I honestly don’t know. But do you? If you think you do, please go read Steve Yegge’s post “Have you ever legalized marijuana?” and come back.

What really bugs me about this whole thing is that people immediately jumped on the fact that Outlook uses Word for rendering, instead of sticking to the real problem that Outlook’s rendering of HTML sucks in some cases. The rendering engine they’re using is immaterial, really, because if the team goes and fixes the rendering inadequacies, then the issue goes away. They can choose how to fix the issues, should they choose to fix them at all, but at this point, even if all the rendering issues are fixed, many people will still be pissed because Word is used to render the email. The argument has shifted from being about the support proper email display to being about Word used for rendering, so there doesn’t seem to be a path to redemption for Microsoft in the court of public opinion that doesn’t involve ripping out Word completely.

I completely agree, personally, that web standards serve as a reasonable basis for email format standards even though there is no formal effort to standardize email. But please argue about the right things. Spend your energy trying to see those standards acknowledged rather than perpetuating this silly argument about ripping out Word from Outlook. Hopefully this effort will have the desired effect, and these rendering issues will get resolved prior to Outlook 2010 RTM. But I can almost completely guarantee that if you want Word completely ripped from Outlook, you’re not going to get what you want.

By the way, does anyone have a screenshot of what the email used in the example image looks like in Outlook 2007, which also used Word for rendering? That struck me as a strange omission. I’m wondering if the issues displayed in that screenshot are just bugs in the 2010 beta. I doubt it, but would still like to see it.

Update: Turns out you can see this in the Email Standards project’s review of Outlook 2007. As I suspected, no real differences.

Also – is there any way to get a sample HTML email from the Email Standards Project’s email acid test? Seems ridiculous there’s no “email me this sample email” form on the acid test page. I can and will file a bug against Outlook if I can get a copy mailed to me.

Update: You can now mail yourself a copy of the email directly from the acid test page.

This didn’t really fit into the overall flow of the post above, but I still think it’s reasonable info to consider, so I’m including it here anyway. I’m about to throw out a bunch of numbers and statistics that are not backed up by any data. They are based only on my own logic and occasionally rational mind. I think they’re true and reasonable statements, but I welcome data that contradicts them.

I think it’s safe to say that a majority, say, 80%, of Outlook users use it with Exchange. Also, a majority of Exchange users use Outlook, and Exchange is primarily used in business settings. Since a large majority of email sent in a business setting is sent to other people in your business, then they’re probably also using Exchange, and also probably using Outlook. Based on this not-so-scientific reasoning, I argue that the number of emails received in Outlook that didn’t originate in Outlook is relatively small. That means, practically speaking, that as long as Outlook can render email that started in Outlook, you’re hitting the majority of your users’ needs.

Now, the idealist in you (and me, for the record) is screaming bloody murder, because you want to see the “right thing” happen for all cases, not just the majority case. But unfortunately, software is more about practicality than idealism, and at some point some smart, but possibly naive people in Outlook made a tradeoff. I’d say with 99% certainty that at some point a developer or two in Outlook estimated the cost of different approaches and implementations, and this one wound up cheaper. They made a cut. They made a tradeoff. And we disagree with the tradeoff.

@biz later posted that they had heard the feedback, claiming it was a scalability problem with the implementation. I initially didn’t buy that argument, but the description over at ReadWriteWeb clarifies things somewhat. This comment over at Winer’s post clarifies it further. I think I understand now, given their architecture, why this would be a beneficial change for them to make.

But that’s not what I’m ultimately writing about. Much of the complaining from people opposed to the change has been that it makes it harder to meet new people on Twitter. If you see a reply to someone else from a person you follow then you might also be interested in following them. OK, fine, I buy that, but for me, I simply want to see everything people I follow say, regardless of who they say it to. With clients like Tweetie and Twitterrific 2.0 (iTunes link) able to browse through conversations effortlessly, I can use the tweets from my timeline as a jumping-off point to see the whole conversation which someone I am interested in is involved in. My goal is not to find new people to follow (though that does happen occasionally), but rather, just to see what interesting things people I follow are saying. I don’t care who they’re talking to. It’s not like everything @Lincolnator says to @MajorLB is completely uninteresting to me just because I don’t follow her.

The proposed changes as they stand right now don’t solve my complaint. But it’s early; the smart people at Twitter will figure something out, and hopefully, it’ll solve my scenario and world hunger at the same time. One can hope and dream.

You can see the protests/commentary onTwitter itself by searching for #fixreplies. It’s still the top trending search term as of this writing.

Stop being idiots.

One school of thought I come across quite frequently is that web content whose markup is not well-formed or is missing required attributes or something just fail to render completely, in order to ensure that all content on the web is gorgeous, standards-compliant markup. This ridiculously draconian viewpoint loses sight of the fact that the ultimate goal of delivering content over the web is just that – delivering content. It seems bad form for a browser to just “give up” when markup is badly formed, because the end-goal of the person building the page – and the person consuming it – is to deliver content. Much of this debate has been chronicled by the IE team; they have a tough job – bring the standards compliance of IE into this century without breaking their customers/users pages. Hence compatibility mode, legacy rendering, etc. etc.

In the past, I’ve always heard this argument from the standards-compliance standpoint. For example, if a page claims to be XHTML but isn’t fully compliant, it should fail to render in a browser. No “best-effort” rendering, just fail. This of course ignores the fact that even the W3C can’t create a parser that can completely validate a page against the spec, but that’s a rant for another time… Assuming the browser can detect that a page is non-compliant, it should just stop.

Anyway, this is a long and winding intro to a post Mark Pilgrim wrote talking about this viewpoint as it applies to accessibility. I had never heard these arguments before, but apparently they’re out there. A choice quote from Mark’s rebuttal (emphasis mine):

I think it would be wise for people who truly care about accessibility to take a closer look at the so-called “experts” who are participating on their behalf, and to understand exactly what these people are proposing. It’s true that some of their proposals have not been adopted, but it’s not because some cartoonishly monocled villain enjoys being mean to them. It’s because the proposals are insane.

Agreed.

There are two core arguments in the article:

1. TV shows are written with commercials specifically in mind, and removing them completely interrupts the pacing of the show.
2. Viewers actually like shows with ads better.

For argument number one, it’s obvious that writers take commercial breaks into consideration when writing TV shows. However, in their DVD counterparts, the breaks are preserved as scene changes. These can seem rather abrupt, and I think this points to an alternative solution: slightly longer breaks for scene changes that are replacing commercial breaks. If we assume the normal scene change is a second, then imagine that the commercial-break-replacement scene change is three seconds. You can simulate this suggestion by hitting pause for three seconds – I think you’ll be surprised at the results. I think you’ll agree that this is uniformly better than ads.

For argument number two, I can’t refute the research. But there is a quote from the article that is completely asinine in my opinion:

From an audience member’s perspective, they are what makes network television social. We use the commercial breaks to talk amongst ourselves, to take bets on the J.D./Elliot situation and to decide that no one ever really dies on ‘”Lost.”

Say what? That we need commercial breaks to be social is ridiculous. In this day and age, we have ready access to the Pause button, and it’s a far more useful tool for fostering discussion while watching TV. Elizabeth and I have been watching quite a few documentaries lately, and we pause quite often to talk about whether or not we agree with what’s being said. I agree that adding discussion and dialog while viewing a show is positive (there’s a reason fans get together to watch episodes of their favorite shows), but I think people use the commercial breaks because they’re there, not because they need them in order to have that conversation successfully.

I hate ads in all forms (though I can appreciate particularly clever or well-done ones), so I am definitely biased here. But seriously… there’s a reason DVD copies of shows sell so well, and it isn’t because people are dying to own season one of “Scrubs” for eternity…

Homer: Not a bear in sight. The "Bear Patrol" is working like a charm!

Lisa: That’s specious reasoning, Dad.

Homer: [uncomprehendingly] Thanks, honey.

Lisa: By your logic, I could claim that this rock keeps tigers away.

Homer: Hmm. How does it work?

Lisa: It doesn’t work. (pause) It’s just a stupid rock!

Homer: Uh-huh.

Lisa: But I don’t see any tigers around, do you?

Homer: (pause) Lisa, I want to buy your rock.

Please please please stop assuming that determining a cause and effect relationship is as straightforward as observing a correlation. It’s not.

PS. I know the != operator is not equivalent to "does not imply." But I am too lazy right now to see if there’s even a Unicode character for the right symbol.

The latest security craze seems to be these security questions. "Mother’s maiden name" apparently doesn’t cut it anymore. Security questions drive me insane, because there’s invariably a finite set of options I have to choose from. Favorite childhood superhero?
Name of firstborn child? What is this crap? None of it is easy for me to remember! Why don’t you let me pick my own question and my own answer? Now I have to remember some "fact" that I made up as an arbitrary answer to some stupid question that a bank decided was an excellent way to distinguish me from some sorry thief.

To add insult to injury, many times I have to select two or more security questions. OMG. I hate banks so much.

When faced with these types of choices, I often forget that one should always pick the low-bandwidth option, even if one is using a high-bandwidth connection. The main reason is that these Flash sites often have annoying background music or annoying animations. From a site design perspective, Flash is best used as a supplementary web technology. A website shouldn’t be designed entirely in Flash. Some core reasons for this are accessibility and search-ability. Google and other spiders cannot crawl a flash web site as easily. Not to mention the fact that these types of sites always have some crazy navigation scheme and flashy transitions between sections that make it very difficult to navigate around, because it doesn’t fit the typical web mold. Finally, Flash sites are typically absolutely scaled at a specific resolution, which means they don’t look or work great if they’re not at a specific window size. To combat this, they’ll pop open a separate browser window at the specific size they want. Talk about annoying! There is very rarely any need at all to design your site in Flash completely. Many Flash-only site have a stripped down HTML version of the site, hence the intro page where you pick which version you want to see. Since you already have to create an HTML version of the site, why not scrap the flash site and focus your attention on making the HTML site much better?

The trend of completely Flash-based web sites has declined dramatically in the commercial web of late, which is a great development. Instead, sites are using Flash as a way to add a bit of pizzazz to the site in strategic places, or to add a slightly more interactive portion to their site. Judicious use of Flash in this way is good.

The next time someone tells you that you should design your entire web presence in Flash, or have two versions of your site, shoot them.

But as I browsed the seemingly endless toothpaste section at my local Fred Meyer, looking for a tube, I realized that I have never bought the same type of toothpaste twice. I get to the store and I realize that I don’t know what type I bought last time, and all of the tubes look the same and have frustratingly similar names. Did I get Mint Zing last time? Or was it Fresh Mint? Maybe it was Minty Fresh… Yeah, it was Minty Fresh.

Nope, it wasn’t Minty Fresh. It wasn’t even Fresh Mint or Mint Zing. It was Mega Mint with Scope, Baking Soda and Peroxide Whitening ++. Sigh. Why does it have to be so hard to buy freaking toothpaste?