If you checked my blog this morning (Boston time) you might have thought your browsing was flaky because columns were strewn hither and yon. Then, if you bothered to reload, they’d be re-strewn in random ways. Sorry. I was working out the kinks in my CSS layout.

But I’ve finally gotten it just about exactly where I want it: Three columns of equal length, with the middle resizable, and colors flush from one margin to another. Plus it seems to work as well in IE as in Firefox. It took many many many hours of work and the unstinting help of some generous readers, but it all came together when I discovered an obscure CSS capability. It’s called a table.

Ok, so I’m being a jerk. Using a table – an information structure – as a layout tool is the anti-CSS. But after breaking this page pretty badly this morning, and after losing another couple of hours of work time just trying to get it back to basic functionality, I gave in and gave up on CSS. For now. Despite the truly generous help of readers, I could not get it to work. The closest I came was to have a pretty good version in Firefox that, nevertheless, had infinite columns running past the footer. And I never got it to work simultaneously in IE, despite trying untold combinations.

I do understand why CSS is the right way to go. I spent most of the late ’80s proselytizing for Interleaf, an innovator in structured editing and declarative markup. I wrote a book on using Lisp to manage structured documents. I was one of the founders of SGML Open (now OASIS). I wouldn’t be caught dead hitting the Enter key twice to separate paragraphs in Word. I’m no expert, but I grasp the basic argument and believe in it. But I couldn’t get CSS to do what I want. It defeated me, to my embarrassment. And if someone wants to do the job right for me, I will not only thank them, I will give her a gift or donate more than that to a charity we agree on. (Talk to me beforehand so we agree about how much I’m paying.)

Here’s what I mean by “right”:

3 column layout, with columns of equal length that size to the longest column. (Compromise: The middle column inevitably is the longest.)

A header and footer, centered

The columns are colored as on this page and touch at the margins The center column’s width resizes when the user resizes the window

The “table” is centered on the page 8px padding

The user can swap in 4 different stylesheets (as per the “Colorblind” option on my home page) The user can toggle the sidebars on and off

The styles are as on my home page

Oversize elements in the center column don’t break the layout

It works in IE, Firefox, Opera, et al.

That is, “works” means that it looks like my current home page in the standard browsers and doesn’t break under normal use.

Until then, I am beaten and humiliated.

11 Comments »

Does Joho look today a lot like it looked yesterday? Are your eyes still burning from the greatest concentration of orange outside of the MinuteMaid headquarters? Still headachey from reading posts so narrow your eyes have to trace a zigzag pattern with the turning radius of a 10th Century spiral staircase? Then I’ve succeeded!

Yes, today’s Joho is a faithful simulation of the original, now done in 98% pure CSS! Gone are the nested tables within tables within tables within tables and the “enough exceptions make a rule” mentality.

Now let the breaking of a formerly stable page begin! (And thank you for all the help! You’re a nice bunch of folks.) [Tags: css]

Allow me to kick off the growsing. If you click on the “archives” or “Search” link right below where the home page page asks you if you’re color blind, in FireFox you get taken to the proper spot on the page but everything above that is cut off. Gone. Not easily brought back. I’m using a simple href=#archives. I didn’t see anything on the Web about this. Anyone feel like a little debugging?

I neglected to thank PositionIsEverything where there’s a spiffy interactive CSS generator for the CSS code for a page with equal-height columns. Well-done!

9 Comments »

I use a bunch of CSS on this page already, but I’m preparing to make the big jump from tables-based layout to using CSS to do the job. Take a look at the source HTML if you want to see why. It’s a freaking mess. Or run this page through a validator. It’s hopeless.

Making the small change required to make this site more readable on mobile devices requires making huge changes. So, I’ve spent a few hours translating this page into a non-table-based, CSS-based layout. Thanks to massive copying of templates (thank you, Web!), I’ve got something up and running that seems to work. (One enhancement I can’t figure out how to do: Have the left and right boxes be fixed in their size while allowing the center one to resize as the user resizes the window. I spent a few hours last night failing to get that to work.)

The new page looks like this one, but the cleanup behind the scenes is massive. But I haven’t posted it yet because I’m sure it’s going to break every way a page can break. I’ll try it later tonight, I think, while some of the US time zones are asleep. [Tags: css html]

I’m trying to get my new pgae to pass the w3c xhtml validator I’m down to a mere 15 infractions. They are of two sorts. First, it doesn’t like Movable Type’s tags. Is there a “loose” DTD somewhere I should be using? If so, what is the exact and precise syntax? Second, it doesn’t like dollar signs and spaces embedded in name tags, as MT insists. So, it flags this line:

<a name=”<$MTEntryID pad=” id=”<$MTEntryID pad=”1″$>

Suggestions? Or should I just learn to love being an outlaw?

7 Comments »

Now that I’ve gotten a Blackberry — a satisfying device so far — I’ve found just what a drag it is to view JOHO on a tiny screen. To get to the blog posts, you have to scroll through an eternity of left column stuff.

Is there a way in HTML to flag which element should display first on a device? I’d like to be able to specify that the top of the middle column is the best place to start if you’re on looking at this page on a tiny screen.

Yes, I could put a name tag such as “start_here” where I want and load the url www.hyperorg.com/blogger/index.html#start_here. But everyone else would be stuck scrolling through from the beginning when they probably really came for the middle column. So, is there a general solution to this problem? [Tags: html mobiles blackberry]

19 Comments »

I have am undoubtedly dumb question about the Semantic Web.

Let’s say I want to express in an RDF triple not simply that A relates to B, but the degree of A’s relationship to B. E.g.:

Bill is 85% committed to Mary The tint of paint called Purple Dawn is 30% red

Frenchie is 75% likely to beat Lefty Niagara Falls is 80% in Canada

Other than making up a set of 100 different relationships (e.g., “is in 1%,” “is in 2%,” etc.), how can that crucial bit of metadata about the relationship be captured in RDF?

Note that I am not asking for practical reasons. I have a theoretical interest in the topic. [Tags: semantic_web rdf dumb_questions]

10 Comments »

In my continuing battle against comment spam, I installed Chip Rosenthal’s Force-Preview plugin. But I’m not having much luck with it, so undoubtedly I’ve done it wrong.

The plugin requires commenters to preview their comments before posting them, a step that apparently spambots can’t take. If the bot tries to post straight through, MovableType rejects the comment because it doesn’t have a token that only gets passed from the preview page.

Except I’m still getting lots of comment spam. I’ve read the ReadMe carefully and believe I have installed it correctly. But there is one anomaly: When I click the “enable” checkbox on the Registered Plugins page and press the “save changes” button, the next time I check, the enable checkbox is unchecked. I don’t have that problem with the other plugins configured via that page.

Anybody have any ideas about what I’m doing wrong, other than using a PC, not exercising enough, and supporting losing Democratic candidates?? [Tags: movabletype plugins chip_rosenthal]

Chip figured it out. I was turning the plugin on at the top level, rather than at the blog-specific level. D’oh! (Thanks, Chip!)

7 Comments »

Roger Dingledine of the Free Haven Project is giving a lunchtime talk about The Onion Router (TOR), an anonymizing router. It’s open source, of course. “We probably have hundreds of thousands of users, although it’s a little hard to tell because we’re an anonymity system.” It’s mainly used by people in Western countries who don’t want to be traced by advertisers, as opposed to being used by dissidents in China or Iran. [As always, I’m paraphrasing and occasionally guessing.]

Encryption doesn’t mask how much you say, when you say it, etc. So, more anonymity is needed. Anonymity is important for privacy, network security and traffic-analysis resistant networks…three ways of spinning the same idea. Citizens, businesses and government need anonymity. E.g., the CIA’s anonymous tip line encrypts the tip but the fact that you submitted a tip can still be traced. “How much would you bid for a list of IP addresses in Baghdad that’s getting messages from the CIA?” By putting all of these users onto the same server, the fact that they’re using it tells you little.

Official Google policy is that they don’t collect personal info. They “only” collect your IP address and what you do with it. But that’s way too much, Roger implies. In fact, he says, if you can see both sides of a Net transaction such as email, you can match up the IP addresses or the contents and make good, practical guesses about who’s talking.

Ethan Zuckerman points out that if you’re the only person using TOR in, say, Sudan, you can probably be identified. Roger says that that’s not something it’d be easy to fix in TOR.

Commercial anonymizers generally put up an anonymous proxy relay. But the user’s request for the relay to fetch a particular Web site could be intercepted. So, some anonymizers encrypt the request. Those are subject to hackers, internal traitors, and legal attacks. The TOR system uses three proxies. (Roger says since the attacks come at the end points, it probably doesn’t matter how many beyond three are added.) If one is compromised, you still can’t connect person A and B. If two are compromised, you can. “We multiplex the circuits because the multiple keys can be pretty slow.” [I record this for your enlightenment. Means nothing to me.] TOR anonymizes only TCP streams. “It needs other applications to clean high-level protocols.”

Server operators are given options to limit bandwidth and choose which ports to connect.

How do you know that a TOR server isn’t compromised, phishing for pigeons? [That mixed metaphor is mine, not Roger’s.] A directory enables servers to vouch for other servers. [Sorry, I didn’t understand that, so I may be misrepresenting it.]

He explains how TOR can provide bidirectional anonymity.

There are about 450 TOR servers and about 200,000 people using TOR in a week. “We push 50MB/second of traffic.”

Problem: “Abusive users get the whole network blocked.” Slashdot and Wikipedia block all postings from TOR. Wikipedia wants to be able to ban abusers’ IP addresses, but TOR IPs are too easy to get. “We make it easy to identify if you’re coming from the TOR network.” They do this on purpose so sites can choose what they want. That means that China, for example, can block the entire TOR network; all it has to do is grab the public list of TOR servers and black them all. To get around this, TOR could have more exit nodes, i.e., last hops from the TOR network that are not recognizable as TOR servers. Roger suggests TOR clients could have a “Help China” button that allows users to forward a small amount of traffic so there would be hundreds of thousands of IP addresses, not 450 TOR servers that are easy to identify and block. He discusses an approach that requires having a trusted social network that grants access to the network.

Next steps: “We want to work on usability.” None of the TOR developers use Windows. Also, incentives: “I really want to do a tit for tat scheme where you don’t get good service unless you handle some traffic.”

Roger mentions that there’s a list of open research questions on the TOR site.

Who are the people who need this, to make it clear that anonymity is good? It can’t be dissidents at this point because, (Ethan says), “It’s a good way to get them arrested.”

Q: What’s the latency of using the system?
A: It depends. But a lot. “If you’re used to university bandwidth, you’ll notice a huge hit.”

Q: Do you throttle people doing video downloads?
A: No. That’s an arms race I don’t want to get into. We ask you not to do huge transfers over TOR. And TOR isn’t very good at that. In some sense, it’s self-correcting.

Q: How are you going to encourage more servers?
A: We have a sign-up on our home page. More important, we’re working on an incentive system.

Q: Do ISPs let people run TOR servers?
A: Many do. By the way, it’s a safe harbor under the DMCA because you’re just passing the bits through.

Q: Some policy makers think we need an accountable network in which we can tie bits back to particular humans. How is TOR going to play in this?
A: Many machines on the Internet are not their “owners” — they do not represent who they claim to. That’s a problem for an accountable network. E.g., Windows 98 is highly vulnerable. Phishing isn’t going away. Roger cites a study that showed that if you add personal info to a phishing attempt — info found by googling — the response rate goes from 15% to 85%.

Q: What will it take for this to move beyond geeky early adopters?
A: There are millions people who’d like to use it, if we can make it easy enough, get it to work well on Windows, get the documentation right…etc.

Q: How can you solve the problem of the lone TOR user in the Sudan?
A: We would need to make it impossible to know that you’re using TOR. You need a way for the data you publish to be unlinkable with you, and defending against that depends on the nature of the attack. There’s also the ability to link up without giving yourself away by someone watching the timing of posts.

A: More users in Sudan would help. Without that, you should go to an Internet site, send a msg to a friend in a safe environment, and ask her to post it for you.

[Tags: berkman anonymity tor roger+dingledine security]

3 Comments »

Jay Allen at Movable Type told me why I was getting this error message when I tried to post this afternoon:

Can’t locate XML/RSS/LP.pm in @INC (@INC contains:

> /path/to/my/MT/install/plugins/spamlookup/lib

> /path/to/my/MT/install/extlib lib

> /usr/lib/perl5/5.8.5/i386-linux-thread-multi /usr/lib/perl5/5.8.5

> /usr/lib/perl5/site_perl/5.8.5/i386-linux-thread-multi

> /usr/lib/perl5/site_perl/5.8.4/i386-linux-thread-multi

> /usr/lib/perl5/site_perl/5.8.3/i386-linux-thread-multi

> /usr/lib/perl5/site_perl/5.8.2/i386-linux-thread-multi

> /usr/lib/perl5/site_perl/5.8.1/i386-linux-thread-multi

> /usr/lib/perl5/site_perl/5.8.0/i386-linux-thread-multi

> /usr/lib/perl5/site_perl/5.8.5 /usr/lib/perl5/site_perl/5.8.4

> /usr/lib/perl5/site_perl/5.8.3 /usr/lib/perl5/site_perl/5.8.2

> /usr/lib/perl5/site_perl/5.8.1 /usr/lib/perl5/site_perl/5.8.0

> /usr/lib/perl5/site_perl

> /usr/lib/perl5/vendor_perl/5.8.5/i386-linux-thread-multi

> /usr/lib/perl5/vendor_perl/5.8.4/i386-linux-thread-multi

> /usr/lib/perl5/vendor_perl/5.8.3/i386-linux-thread-multi

> /usr/lib/perl5/vendor_perl/5.8.2/i386-linux-thread-multi

> /usr/lib/perl5/vendor_perl/5.8.1/i386-linux-thread-multi

> /usr/lib/perl5/vendor_perl/5.8.0/i386-linux-thread-multi

> /usr/lib/perl5/vendor_perl/5.8.5 /usr/lib/perl5/vendor_perl/5.8.4

> /usr/lib/perl5/vendor_perl/5.8.3 /usr/lib/perl5/vendor_perl/5.8.2

> /usr/lib/perl5/vendor_perl/5.8.1 /usr/lib/perl5/vendor_perl/5.8.0

> /usr/lib/perl5/vendor_perl .) at

> /path/to/my/MT/install/plugins/mt-rssfeed.pl line 107

Jay says the last line tells you which of your Movable Type plugins has gone south. In my case, it was mt-rssfeed. All you have to do to solve the problem is go to the plug-ins configuration panel and disable the plugin. Or re-install the plug-in.

Be the first to comment »

The Google Pack of free software consists of packages I’ve either already installed or already uninstalled. It’s way more boring than what we’re used to from Google. But ZDnet UK has an interesting take on it:

It’s an upgrade channel, a path straight to your desktop. Microsoft has one with Windows Update; Google wants the same. Expect this to develop rapidly, with many more options such as Open Office — too big, too many implications for the first release — and plenty of pre-installs on new computers.

So long as Google doesn’t do what Microsoft does with its automatic upgrade facility: After installing a critical update, it sometimes reboots your machine without asking you. (It’s a settable default but it’s set to default the wrong way.) [Tags: google googlePack microsoft]

3 Comments »

Yahoo has done a good job emulating the Mac’s Dashboard widgets with its own Yahoo Widgets. (Yahoo bought Konfabulator to get its Widget engine.) They even look cool like the Mac ones do. But, Yahoo is not doing a good job helping us to find them. For example, there are 186 widgets in the System Utilities folder on the Web site, and 276 in the Fun and Games category. Then there are nine more categories. That’s too much to browse, even if you set it to show 25 at a time.

How about taking a page from Firefox Extensions and offer listings by most popular and highest rated? Or, heaven forbid, do we need a meta-Widget? [Tags: yahoo widgets macintosh firefox]

8 Comments »