July 26, 2004

I'm looking over...

Posted by Scott at 10:16 PM

Clover lawn - After over a year of indecision, we decided to overseed the lawn this afternoon with 10 pounds of Dutch white clover seed. Clover?! Isn't that the stuff most folks are trying to get out of their lawns?

We've been doing a lot of research on lawns, especially where dry conditions or poor soil exists. We live in the 'Granite Town' of the 'Granite State'. This is definitely *not* the Chicago area, home to some of the best topsoil around! When digging on our lot, it isn't uncommon to hear the shovel go "clink", even just a few inches below the surface. Last year Adelphia was laying some new underground cable along the street. I couldn't believe the number of rocks that they pulled just to lay a cable! I almost could have built a traditional New England stone fence. Even though we have in-ground sprinklers (including *50* heads), if the weather gets hot and dry for a few days in a row, the lawn starts to turn brown.

In doing this research, a theme that kept creeping back up was to intentionally blend clover into your lawn. The article that most swayed me was this one at about.com. It was backed up by this, this, this, and this. Overall we're hoping it can stay green better than the current monoculture of grass that the likes of TruGreen advocates. What I found interesting is that clover actually puts nitrogen into the soil whereas grass depletes it, requiring frequent fertilization to keep up. Clover also tends to have lower watering requirements (important to those of us whose water comes from a well!) and aerates the soil with it's root structure. The important thing will be to mow the flower heads when they get too dense so as not to attract too many honey bees. I don't want the girls freaking out.

When the package of seed arrived late this morning, Michelle still had a moment of hesitation. I don't blame her. It's a rather bold move that goes against everything TruGreen/Chemlawn and LawnDawg tend to stand for. We were walking down the street yesterday with the kids. A neighbor a couple of doors down had clover blended in with his grass. It was thick and green without any sprinkler system and looked nice when maintained and trimmed. Because it was uniformly spread throughout, it didn't seem patchy. We'll see how it goes. We're keeping our fingers crossed.

MovableType - Warning! Lengthy, geeky discussion follows. Click only if you care to read it. Otherwise move on to the next post...

A few months ago there was a lot of whining going on in the weblogging community because Six Apart (the company that makes the software this site uses) announced a developer's version (3.0) of MovableType (the software this uses). In versions prior to 3.0, such as the 2.661 that I'm on, the software was free for personal use. The newer 3.x series was going to have fees if it went outside a 'personal' usage model as defined by Six Apart. Initially they weren't careful in their language, but later clarified what counted as free for personal use. For me it means that because Suzy and Claire are co-authors, there is a nominal fee. I can live with that. I still consider MovableType to be great weblogging software with a huge base of fellow users who are always discovering innovative things to do with it. Today Six Apart gave a little more insight into what will be in the general (in other words, not just for developers) release of 3.1, scheduled for a few weeks from now. Once it's ready, I plan to pay up and download.

I'm one of the few people who really likes what some consider MovableType's weakness: it generates static files for all weblog data. I tend to strongly believe that this is sufficient for 99.9% of users who would want to write a weblog. As someone involved in web technologies for my company, I know that dynamically generated websites have their place. Amazon, eBay, Yahoo, etc. I just tend to think the dynamic approach is overkill for most personal websites. Compare what happens in the two approaches. When you hit the homepage of this site, your browser makes a request, the webserver finds an appropriate file sitting on it's local hard drive (or better yet, in cache) and serves it up. Brain dead simple and efficient. In most dynamic weblogging software the server uses an interpreter that evaluates a program. This program almost always makes several network requests to a separate dedicated database server to extract the last several entries. From there it stitches those entries together and sends them back to your browser. Typically it only takes less than a second, which wouldn't be bad if the web server was only hosting your site, but often those webhosts are packing in as many users as they can onto each server.

Now consider someone who comes back to your site to see if anything new has been added. In the case of static files, the server checks the timestamp on the file and might just effective reply "nahh, nothing new here". It's a quick, easy check for the server and the response takes almost no network bandwidth. Even 56k phone modem users get a quick response. When your browser sees this response, it just quickly pulls up the last copy from your hard drive. This is why when you go back and look over last week's photo album again, it's like lightening. In the case of a dynamic site, in almost all cases, the webserver doesn't know. It almost never has any kind of "freshness" tag so it usually just starts all over. A few with special coding will actually query the database and eventually come back to say, "nahh, nothing new here" but most dynamic sites go for the lazier approach and just recompute and resend the page every time.

What's even more absurd to me is that most weblogs only change a few times per week. Everytime anyone goes to that page at a dynamic site, it always does the same interpret/database query/page stitching operations. Over and over and over. What a waste! It's amazing that many computer owners know that cache memory is a valuable thing in a computer but most semi-tech savvy webloggers put up with this constant re-computation of their webpages. That's what caches do: they optimize for the common case by making frequent accesses easy to reuse.

There are certainly some things that can only be done with dynamic approaches. One example is the new fad of timestamping entries with relative time rather than fixed time. In other words using "4 hours ago" rather than "5:30pm". There are other cute tricks that can only be done with "just in time" computation, but for the most part, it's just a weblog. There are better ways to show how clever (read: geeky) you are.

I say this because I just bought a book for work on PHP, the most common web scripting language on the internet. The sites we will setup at work will have a strong dynamic component with frequent ties to various databases. I had considered moving BilikFamily.com to alternative dynamic content weblogging systems based on PHP such as WordPress and Nucleus, as they are the leaders in that paradigm. I still may but primarily for the practice in using PHP. I still find the approach overkill for the random rantings of Joe Q. Public in a weblog. If a site like TownHall can manage to archive it's dozens of new daily articles primarily using static files (albeit with server side includes for the advertisements), I think most webloggers should be able to as well. Focus on content, ease of use, and quick page refreshes, not your latest "show what song I'm listening to right now" plugin/extension!

Comments