Long time, no blog. Anyhoo, I had Amanda download Adobe Photoshop Album and try to use it, as she was jealous of what I could do in iPhoto. Well, after she downloaded she spent about 6 hours trying to do what she wanted, and it would not cooperate for her. During the same time period, I was going through the Adobe Photoshop Classroom in a Book (a little slow, but good nonetheless), and came upon lesson 3. Lesson 3 is a lovely little lesson wherby you crop, adjust color, etc. It was very useful for a deadbeat left brain like me. After showing Amanda what I could do in Photoshop, she wanted all her pictures doctored ala lesson 3. So I spent a good deal of time preparing pages to print of various photos from Evie’s retirement. It was quite fun, actually, but Photoshop will suck time faster than a black hole. I can see now why Photoshop is Alex’s choice when it comes to graphics work.

But alas, the weekend is over, and the left brain has to come back to the foreground to make the dough

Popularity: 10%

Just experimented with the joys of WebDAV and remote editing. I put my website under source control last night, not unlike the way we do it at Apache. Then I got to thinking, that deploy step is just a hassle, why can’t I just edit the content on the server? Come to find out, I can. This is great. The following is a mini-HOWTO on WebDAV with Apache2 and Mac OS X:

Edit your httpd.conf to include the mod_dav (WebDAV) and mod_dav_fs (filesystem based store) modules:

Then we also add the mod_dav_fs lock database statement:

Make sure to set this to a drectory (/www/var here) that is writable by the user AND group of the Apache process (nobody:nobody for me).

Then we go into some website config, and add the magic incantations that export the website as editable:

Now for some explanation. Since I have static and dynamic content on my site (who wouldn’t), I have aliased the ‘/source’ URI so I can edit both types of content. The ‘Dav On’ statement tells the mod_dav module to get to work on this directory. To prevent dynamic content from being executed, the ForceType directive tells Apache to just spit out the raw conent (text) of the files. This means index.php will be the source instead of the result. The Auth* directives say that I want to authenticate the users that access this resource, as does the Require directive.

So now we have Apache setup correctly, we restart it (apachectl restart), and go to test it. How?, you ask. Simple: Jaguar. Mac OS X (I am using Jaguar 10.2.6) has WebDAV filesystem connections included in the OS (this is how they connect to your iDisk on your .mac account)! This is going to be too easy. Just open a finder window, click command+K (Connect), type http://dotnot.org/source (your URL will be different of course) into the address box, click connect, and we are done. On the Mac, we find this info under /Volumes/source, and we can just copy, delete and edit files in there, all the time editing the remote content. This is too cool.

A few comments before you go crazy remote-enabling everything, a couple of warnings: First, this method is only as fast as the network connection it is operating on. So mass updates should probably be done locally, and batched up. Second, you are editing the live content. This means there is no backup, unless you make one. I would highly reccomend using some soure control product, such as Subversion (which uses WebDAV by the way) to keep archives of the files. And finally, this should work on Windows as well (using WebFolders and IIS), but that is not my setup. I would be happy to post similar information on how to do this on Windows.

Feedback appreciated.

Popularity: 11%

Mostly painless. Updated to latest Apache (2.0.47), Subversion (0.25), and PHP (4.3.2). Also upgraded tasks to 1.6.1.

Popularity: 19%

I have googled most of one day to find this, to no avail. I am running two apps in OS X that need access to environment variables. The specific apps are Intellij IDEA 3.0.4, and Emacs 21.3. The specific environment variables that I need both of these apps to know about are CVSROOT and CVS_RSH. I access CVS over ssh, so I have set the CVS_RSH env var to ’ssh’ in my tsch shell init file. In the Terminal, everything works great. In Emacs, I get a connection refused. I have tracked this down to Emacs attempting to contact my CVS server over rsh, which is the default.

How do I tell Emacs and IntelliJ about CVS_RSH=ssh?

I am willing to pay as much as $20 for the person who can get both IntelliJ and Emacs (my versions on my machine, of course) to work properly.

Popularity: 12%

We need a Rural Broadband Act, quite like the Rural Electrification Act (A better explanation of the act is here. The internet is the ‘library of tomorrow’, the new communication pipe, and it exists here today in urban areas. We need to make it exist for every person in this country, if we would like to keep any edge we may have. Nost just dial-up access. I am talking true broadband, minimum 1Mbps but more like 100Mbps considering the time it would take this bill to pass. The conomy is in the dumps, this would create new jobs, not only on the infrastructure front, but on the internet front as well.

But wait, Google says the Rural Broadband Enhancement Act (RBEA) has already been attempted. Well, what happened then? If we look at this page, we see that everything regarding rural broadband deployment was refered to committee in early 2000, and is not out yet. Will they ever come out of committee? How long does something sit in committee?

This research was pretty dishearening, and I think that the blogging world might do some good by latching onto this and attempting to move it from its existing glacial path.

But I could be wrong…

Popularity: 12%

Ted Leung has an excellent thought If RSS aggregators are TiVo for blogs.

I couldn’t agree more. You’re RSS feed shouldn’t be some teaser to get me to your site. Your are locking me into the browser age. Publish all of your content in your RSS feed, and your entire voice will be heard, not just the sales pitch. Why are you trying to get me to come to your site in the first place? To see your site design? Or maybe it’s the Google AdSense? Let’s make RSS aggregators first level internet tools, equal with, not subbordinating to, the browser.

Just testing the threshold, JT

Popularity: 11%

In Information Foraging: Why Google Makes People Leave Your Site Faster, Jakob Neilsen compares web browsing with animals hunting in the wild:

This is one article that I can completely agree with. In watching myself, Jason, Alex, even novice web users, I see all of these cues. Even my mother in law will become frustrated if she can’t find what she is looking for. Her ‘preferred’ hunting ground is the bookstore, in the magazine section, so she will revert to that. My only hope is that Google and others do get better. The concept of a sticky site should come from the worth it provides to the community, not from flashy looks, or preventing the back button, etc. Here at dotnot, I don’t even have anything on my home page yet, because I am resisting posting anything that is non-essential. I worked hard to make the HTML of my site simple, I want the content to be just as simple, and convey the necessary information without all of the fluff.

And here is a definition:

Popularity: 13%

The dotnot site is now under source control, with the help of Subversion. Let me know if anything seems weird (permissions problems?), etc.

Alex and Jason, I accidentally delete the .htpasswd file for tasks, so we need to do that again.

Now that it’s under source control, now I can start adding the content. Step 3: profit!

Popularity: 19%

Released Jaxen 1.1 beta 2 yesterday, so XMLC 2.2 can be released. New features: HTMLXPath implementation that can convert to upper-or-lower-case on xpaths, and SaxPath being rolled into the Jaxen codebase.

The major bug fixed is the ability to register extension functions in a namespace.

1.1 beta 3 should be coming up in the next couple of weeks, and will switch to support JDOM 1.0beta9 (they f#*king changed their API), and more bugfixes. Come on over and join the party.

For a little bit of history, I have started work on Jaxen because my current employer is doing dynamic xpath evaulation in Saxon, which while very fast in most areas, is very slow in that one. We are switching to a SAX pipelined filter architecture, where the stylesheet just renames elements (XSLT is good for that), and using Jaxen in the filters for the dynamic evaluation. In tests we did, Jaxen on DOM (0.07ms) was 7x faster than Saxon (0.5ms) in dynamic xpath evaluation. When you have a few thousand of these to do, it really adds up. I am not posting total results, because it was not intended to be a benchmark. We tested our application, in our environment, with our data, and decided that Jaxen fits the situation better. I would encourage anyone to do the exact same thing when looking at A vs B vs C.

My future with Jaxen seems to be bright. I am soon to start work on a Navigator for JavaBeans, and then a GenericNavigator to cross any model that Jaxen currently supports. Oh yeah, and maybe include an existing XOM implementation if I get the go ahead.

Popularity: 12%