Archives


Atom Feed

Associated Blogs
 
Edge City
 
Numbering Peano
 
Orcmid's Live Hideout
 
Praxis101
 
Prof. von Clueless in the Blunder Dome

Recent Items
 
Friday Cat Picture: In the Princess Eye (from 2007...
 
Friday Cat Picture: Weez Ur Trickz n Treatz
 
Friday Cat Photo: Princess Psyche (from 2007-09-24...
 
Friday Cat Picture: Streaks at 21 (from 2007-09-07...
 
Geek Dinner Collection: 2007-09-12 Hanselman Event...
 
Confirmable Experience: What a Wideness Gains
 
Confirmable Experience: Consider the Real World
 
Suono: Playing with HDR Photography
 
Interoperability: The IE 8.0 Disruption
 
Golden Geek: Executives and Malcontents

Interoperability: The IE 8.0 Disruption

I've elected to adopt the IE 8.0 beta 2 release as a tool for checking the compatibility of web and blog pages of mine.  I see how disruptive the change to default standards-mode is going to be and how IE 8.0 is going to assist us.  I need to dig out tools and resources that will help me mitigate the disruption and end up with standards-compliant pages as the default for new pages.

Looking Over IE 8.0 beta 2

I avoid beta releases of desk-top software, including operating systems and browsers.  Because the standards-mode default of IE 8.0 is going to place significant demands on web sites, I also thought it time to install one copy of IE 8.0 simply to begin assessing all of my web sites and blog pages for being standard-compliant enough to get by.  I am willing to risk use of beta-level software in order to be prepared for the official release in this specific case.  I'm also sick of having IE 7.0 hang and crash on mundane pages such as my amazon.com logon.  I'm hoping that even the beta of IE 8.0 will give me some relief from the IE 7.0 unreliability experience.  And so far, so good.

With the promotion of beta2 downloading this past week, I took the plunge.  Installation was uneventful and all of my settings, add-ins, favorites and history were preserved.  My existing home page, default selections, menus and tool bars were also preserved.  [I am using Windows XP SP3 on a Windows Media Center PC purchased in September, 2005.  IE 8.0 beta 2 also seems faster on this system in all of its modes.]

I did not review much of the information available on IE 8.0, expecting to simply try it out. 

My first surprise was a change to the address bar.  There is a new format where all but the domain name of the URL are grayed.  That was distracting for the first few days and it still has me stop and think.  I realized this is the point: emphasizing the domain name so that people will tend to check whether they are where they expect to be.  I like the idea, even though I have to look carefully and remember the full URL is there when I want to paste it somewhere or share the page on FriendFeed or elsewhere.  I take this provision as one of those small details that demonstrates a commitment to safe browsing and confident use of the Internet.

What I was looking for, and saw immediately, is the new compatibility-view button.   This "broken page" button appeared on the first site I visited after installation of IE 8.0 beta 2.

Clicking the button causes it to be shown as depressed and the page is re-rendered as a loosely-standard page with the best-effort presentation and quirks renderings of IE 7.0 and earlier Internet Explorer releases.  If you leave the button selected, the setting is remembered and automatically-selected on your next visits to the same domain.  It stays that way until you unselect the button by clicking it again while visiting pages of that domain.  It was this feature that tipped-me over in wanting to check out my own pages using beta2 (although I thought the button was tracked at the individual page level until I read the description of domain-level setting). 

By the way, if a page is detected to require a standards or compatibility mode specifically, no compatibility view option button is presented.The amazon.com site is this way from my computer, and so is Vicki's pottery-site home page.  I looked at the source of the amazon.com site and confirmed that they are not using the special tag that requests that the compatibility view be automatic.  I didn't check the HTTP headers to see if they are using that approach to forcing a compatibility or a standards-mode view.  I know I did nothing of the kind on Vicki's site.  This suggests to me that there is also some filtering going on in standards-mode rendering to notice whether a compatibility view should be offered.  I'm baffled here.  I am curious whether there is any browser indication when the compatibility view is selected by a web page tag or HTTP header.  I suspect not and I'll have checked into that soon enough.

I also checked out the InPrivate browsing feature, which, although popularly dubbed the "porn mode," is very useful when using a browser from a kiosk or Internet cafe and when making private on-line transactions from home.

At this point, I am not interested in special features of IE 8.0 other than those related to improving the standards-compliant qualities of web pages and the browsing experience.  I may experiment with other features later.  My primary objective is to use the facilities of IE 8.0 and accompanying tools to improve the quality and longevity of my web publications.  Once I have some mastery over web standards, I will look into accessibility considerations, another project I have been avoiding.

Disrupting the State of the Web

The problem that IE 8.0 is intended to help resolve is the abuse of Postel's Law [compatibility view offered] that the web represents: "be conservative in what you do, be liberal in what you accept from others."  The abuse arises when what you do is based on what is being accepted, with no idea what it means to be conservative.  The web was and is an HTML Wild West and it is very difficult to enforce conservatism (that is, strict standards conformance in web-page creation).  Since browsers also varied in what they accepted and then what they did with it, loosely-standard pages and loosely-standard browsers have been the norm and web pages are crafted to match up with the actual response of popular browsers. 

Since Internet Explorer is made the heavy in this story, we now get to see the price of changing over to "be strict in what is accepted and be standard in what is done with it."  This is a very disruptive change.  We'll see how well it works.   Joe Gregorio argues that exceptions to Postel's Law are appropriate.  Some, like Joel Spolski [no compatibility view], think it might be a little too late.  There are already some who claim that the IE 8.0 Compatibility view is a sin against standardization [compatibility view offered], no matter that not many of the 8 billion and climbing pages out there are going to be made strictly-conformant any time soon.  With regard to compatibility mode, I think it is foolish for it not to be there and Mary-Jo Foley is correct to wonder how much complainers are grasping at straws.

It was surprising to me to observe how regularly the compatibility-view option button appears and how terribly much of my material renders in IE 8.0's standards mode.  Apparently the button is there because IE 8.0 can't tell whether the page is really meant to be rendered via standards-mode or is actually a loosely-implemented page.  I'm spending a fair amount of time toggling back and forth to see if there is any difference on sites I visit.  This suggests to me that there is going to be a rude awakening everywhere real soon now.  It is also clear to me that I don't fully understand exactly how this works, and I need to find a way to test the explanation on the IE blog and the discrepancies I notice, especially when the compatibility-view option is not offered and I know nothing special was done to accomplish that on the web page I am visiting.  I am also getting conflicting advice when I use an on-line web-page validator.

This change-over to unforgiving, default-standards-mode browsers is going to be very disruptive for the Internet.  In many cases, especially for older, not-actively-maintained material, the compatibility view is the only way to continue to access the material successfully.  There is a great deal of material for which it is either too expensive or flatly inappropriate to re-format for compatible rendering using strictly-standard features.  Without compatibility view, I don't think a transition to standards mode could be possible.  The feature strikes me as a brilliant approach to a very sticky situation.

Although there is a way to identify individual pages as being loosely-standard and intended for automatic compatibility view, that still means the pages have to be touched and replaced, even to add one line to the <head> element of the HTML page.  There are billions of pages that may require that treatment.  Perhaps many of them will be adjusted.  That will take time.  Meanwhile, having the compatibility-view option and its automatic presentation is very important. 

There is also a way to adjust a web server to provide HTML headers that request a compatibility (or standards-mode only) view of all pages from a given domain.  That strikes me as a desperate option to be used only when there is no intention of repairing pages of the site.  I might do that temporarily, but only while I am preparing for a more-constructive solution that doesn't depend on compatibility view being supported into the indefinite future.  The variations on the available forms of control (browser mode, DOCTYPE, HTTP header, and meta-tag) need to be studied carefully.  I expect there to be confusion for a while, probably because I am feeling confused with the ambiguities in my experience so far.

Another problem, especially with regard to IE 8.0 beta2, is that we don't reliably know how badly a loosely-standard page will render with a final standards-mode browser versus the terrible standards-mode rendering that beta2 sometimes makes at this time.  It is conceivable that the degradation might not be quite so bad as it appears in beta2, but there is no way to tell just yet.

The need for expertise and facility with semi-automated tools as part of preserving sites with standards-conforming web pages is probably a short-term business opportunity.   The web sites that may be able to make the transition most easily may be those like Wikipedia, where the pages are generated from non-HTML source material.  (That makes it surprising that Wikipedia pages currently provoke compatibility buttons and compatibility view is needed to do simple things like be able to follow links in an article's outline.)

Mitigating IE 8.0

To mitigate the impact of IE 8.0 becoming heavily used, it is necessary to find ways to do the least that can possibly work at once, and then to apply that same attitude in making the next most-useful change, and so on, until the desired mix of standards-compliant and loosely-compliant pages is achieved.

To find out what tools are available along with IE8 beta 2, these pages provide some great guidance and resources:

• Ed Bott 2008-08-28: An IE8 Beta 2 Q&A
    
• IE8 Blog 2008-08-27: Internet Explorer 8 Beta 2 Now Available
    
• Microsoft: Windows Internet Explorer 8 (beta): Home Page 
    
• US ISV Developer Evangelism Team: Consumers Begin Using Internet Explorer 8 Beta 2
    
• MSDN online: META Tags and Locking in Future Compatibility (preliminary), with details for how to mark pages and also set the sites HTTP response headers (not for the faint-hearted)

That should point you to all of the resources you need to understand how to check sites, how to use the compatibility provisions, and other ways to take advantage of IE8 availability when it exits beta.

I'm looking at a progression that will allow the following:

1. Have a complete site automatically set to be browsed in compatibility mode (EmulateIE7, in my case), buying time to provide finer grain solutions
        
2. Modify templates on blogs such as this one to specify compatibility mode on all new and updated pages until I say otherwise
     
3. Find a way to make bulk changes to pages, adding a <meta> head element that specifies compatibility mode for those pages
     
4. Decide how to migrate pages so that their results are delivered best in standards mode.  This may be a very long-term approach that doesn't begin implementation until the percentage of old browsers still in use diminishes enough to have standards-mode browsers be dominant.  There should still be a substantial period of time while compatibility mode is grand-fathered by the latest browsers.
    
5. Even if compatibility modes eventually disappear from popular browsers or whatever comes after the browser, there will be a lasting need for compatibility view of archival materials, or some other creative solution that allows those materials to be accessed in a standards-mode world.
     
6. There will be future breaking changes in standards mode as updated/successor standards are introduced.  The compatibility view requirement may never disappear, although its future achievement may be accomplished with less disruption.  Unless, of course, we fail to learn the lesson.

I will work out my own approach on Professor von Clueless, since I have definitely blundered my way into this.

Labels: confirmable experience, IE8.0 mitigation, interoperability, trustworthiness, web site construction, web standards

 
Comments:

 
just make it simpler: use Firefox

you have better things to do that support poor browsers that never took standards serious.

# posted by  franco: 31 August, 2008 15:24

 

 
My concern is not about choice of browser, it is about the level at which my web-site and blog pages are standards-conformant and will render properly with a standards-conformant browsers.

It happens that the IE 8 beta 2 compatibility-view option is giving me a way to confine my incompatibilities and then remove them as browsers all become standards-compliant together.

# posted by  orcmid: 31 August, 2008 17:50

 

Post a Comment

Interoperability: The Experience of It

It strikes me that interoperability is a lower-level technical quality that impacts the higher-level undertakings of people, groups, and organizations.  The connection between interoperability as a technical achievement and how its absence or limitation comes to be noticed is indirect.  I wonder how important it is to sort that out.  Here are my exploratory thoughts.

cross-posted to the Interoperability Forums

Having Interoperability Conversations

In April, on the Interoperability Conversations Forum, I asked:

Who is dropping in here to discuss interoperability?

At what level of detailed concern?  The general level of the Interoperability Principles, the Document Interoperability Initiative, or some narrower or broader (industry level, even) focus.

Also, at what level of "system" is interoperability of interest?  The kind of technical interoperability that programmers deal with when integrating software, the interoperability that IT infrastructure and systems integrators are concerned with, or the broad level of interoperability between individuals and groups within, across, and between enterprises and communities?

The people who mention that forum thread to me echo the reply from N. Gregg Brown:

"I would invert the levels of interest that you gave at the beginning of your note to represent the topics we hope to discuss. First, I’d like to see us discuss the scenarios that people want (are itching) to accomplish so they can collaborate “between individuals and groups within, across, and between enterprises and communities.” From that will flow the issues for IT and Web that providers will have to deal with and, inevitably, specific issue that programmers will have to address. ...

"If we can discuss how these things raise the issue of interoperability, we can make better judgments about the requirements that need to be fulfilled, the standards that need to be adopted, and the stuff that needs to be fixed."

I hadn't meant to suggest a sequence.  I am inclined to begin with how interoperability serves the activities of individuals and groups in their mutual endeavors.  I will go so far as to claim that it is all about serving people in their endeavors and removing those barriers that lack of interoperability represents.  There are some problems with that, though.

How Is Interoperability Experienced?

I don't think people experience interoperability or its absence in those terms.   I think people experience the conceptual integrity of computer-based systems;  they experience the coherence with which the elements of a system are familiarly usable; and, more likely, they experience frustrating absence of those qualities.  And people do not identify and report their experiences in those terms.

I think that what people stumble over is, among other defects, an absence of interoperability.  It shows up as being frustrated in some task that interoperability is expected to enable.

The arrangements for interoperability are subordinate to and invisible at the level where most of us interact with computer-based systems.  The use of standard document formats is an easy example.   When I am creating a document, my attention is not on the the digital representation of the document, it is on what the application allows me to do, how the document is presented to me, and what happens when I and others use the resulting electronic document.   Even when there is a problem exchanging the document with someone else, I don't go digging into the format as a way of resolving the situation.  And for most people, examining the digital format is out of the question.  Digging into the computer software that is involved is more unlikely even if possible.  Most of us are similarly unlikely to examine the communication protocols our computers are employing to connect and interact over the Internet.

I think that interoperability is known to us mostly as an expected check-off item; it is not something we think about.  Unless it doesn't work.  At that point, the breakdown might be identified as an interoperability failure or not. 

Recognizing Layers and Flavors of Interoperability

I do think there are layers of interoperability.  When we talk of interoperability, it is important to be clear what the layer and the context are.  I think we also must be clear about what aspect of interoperability matters.  For example, are we after substitutability in the same layer or are we interested in substitutability above or below a layer.  Is it ever not about substitutability?

I also think that there are interoperability-related limitations that are experienced at a higher level where the (non-) interoperability contribution is not apparent.

For me, this means that I must listen carefully for stories about interoperability.  I must be careful in identifying the different layers or levels that are involved and where the interoperability matters.  There's a lot of trial and error in sorting that out.

The nagging question that remains is: how can interoperability be demonstrated in a way that is relevant at the other levels where its achievement (or failure) will be felt? 

Labels: interoperability

OpenOffice.org: Another Hot Tip!

[update 2008-10-13 The installer bug observed here is not specific to OpenOffice.org, it seems to be specific to installers.  I have not examined this enough to see which installers do this, but I have seen the identical problem with installers of other software.  In all cases, the software will correctly place a single-user install under the user who is running the install.  But the dialog identifies the wrong user, apparently always showing the User ID for the first user created on the machine.]

When I installed OO.o 2.3 on my sister's computer, I was disturbed that it kept offering her admin account as the single account it would install under, even though we were not running the installer under that account. 

At my XP SP2 system at home, I installed the same version and I did not have that problem.  This time it did name the account I was using, even though it was not my normal administrator account.  It was, however, the first account that had been set up on my machine, as was the case for admin on my sister's machine.

So I tried again, this time on my Tablet PC and Windows Vista Ultimate.  For variety, I also used the OO.o 2.1 Novell edition, installing from CD-ROM.  There, I ran into exactly the same problem.  I was presented with this dilemma:

Once again, me is not admin.  I am doing this install from my standard-user account (SUA).  But just to see what would happen, I took that option anyhow.  Guess what: This dialog is lying.  It will install only for the account being used.  The bug is that it doesn't present the correct account name.  The behavior is actually correct.

So if you are attempting to install OpenOffice.org 2.3 (or the 2.1 Novell Edition) only under the account you are running in, you can ignore the incorrect account name.  It will do the right thing. 

The next time I assist my sister in adding an OpenOffice.org update, I'll be sure to uninstall the current version and then install the new one only for her standard account.

Now, you might wonder what the fuss is all about.  If you are as obsessive as I am about computer security, you might want to omit all but pure administrative applications from the administrative account, and only ever use the administrative account for essential administrative operations.

This means that to have ordinary applications install properly in the ordinary accounts where it is safest to run them I elevate my standard-user account to an administrator account just long enough to install the software and run it the first time under the standard account.  This gyration is required because many programs expect to perform final administrative setup operations on the first execution.  Setting of registry entries and creation of application data, plus other details, may be specific to the account that is used for the install.  I will usually discover the firewall conditioning that is required upon the first execution.  From then on, I can use the program as a standard user.

When certain programs (e.g., Second Life) install for all users with no other option, I will remove the shortcuts and links placed on the "All Users" desktop and startup menu and place them in the profile information of my standard user account.  This is just a little preventative against my foolishly using recreational software from my administrative account.

[update 2008-10-13 Moved this post from Orcmid’s Live Hideout to Orcmid’s Lair for better preservation and tie-in to other confirmable-experience and cybergeek topics.]

Labels: confirmable experience, cybersmith, interoperability, software usability