Turns out they’re actually hosted by Bluehost, which offers a very similar hosting package. Hosting discussions seem to be coming up a lot lately in various places I frequent, so I thought I’d share what little I know about finding a good web host.

“Unlimited” != Unlimited

First of all, “Unlimited” hosting packages are not unlimited. Even in the virtual world of computers and the Internet there’s no such thing as limitless resources. In fact, hosts that offer unlimited packages know this, and they have specific disclaimers in their terms of service that repeatedly point this out, though rarely in explicit terms. From the Bluehost TOS:

BlueHost.Com does NOT provide unlimited space for online storage, backups, or archiving of electronic files, documents, log files, etc. …In most cases, a Subscriber’s web site will be able to support as much traffic as the Subscriber can legitimately acquire. However, BlueHost.Com reserves the right to limit processor time, bandwidth, or processes in cases where it is necessary to prevent 

Hosts that offer unlimited packages always have things like this in their Terms of Service, and it basically means they can shut down your site if you start using excessive system resources. Because the plans are described as unlimited, though, what constitutes excessive is left entirely up to the host. This page offers a little more detail on what unlimited really means and when to consider choosing a host that offers unlimited hosting.

Know What You’re Getting

Continuing from the previous point: read the Terms of Service. You’ll be entering into a contract and you’ll need to know what your rights are and what the rights of the host are. Dreamhost, another unlimited provider, has this interesting bit in their Terms of Service:

DreamHost is under no obligation to compensate Customer for downtime, whether the downtime be caused by Customer, DreamHost Web Hosting, or DreamHost’s upstream providers.

And that’s for dedicated hosting at $395/month. Any good webhost should be able to gaurantee something in the are of 99% uptime, with some kind of compensation for excessive downtime (especially if you’re looking at dedicated hosting plans). Like unlimited claims, though, don’t fall for “100% Uptime” claims.

ICDSoft, who Michael Online currently hosts with, doesn’t have anything like that in their terms of service but does explicitly forbid all kinds of things on their servers, including porn, warez, running IRC servers, proxy software, etc. Read the Terms of Service.

Limits and Support are your Friends

If you look at ICDSoft’s packages, you’ll notice that they look almost meager compared to what Bluehost or Dreamhost offer. What they do offer is top-notch service. One thing to always keep in mind is that your entire website is basically at the mercy of the host and their servers; if something goes wrong, you’ll be going to them for help. Look for quick turnaround on support requests and check to see how helpful the host actually is.

Like good support, limits are actually worth paying a little more for. ICDSoft doesn’t offer “unlimited” packages, but they do offer explicit limits, so you know exactly what you have to work with and what will happen if you go over those limits. For a higher-end example, however, take a look at MediaTemple, which offers a selection of hosting packages (none of which offers “unlimited” anything). Notice the complete lack of “Unlimited!” claims, and the rather high cost of hosting: their basic plan starts at $20/month, and plans can go as high as $750/month.

Pay for What You Need

The most important thing about finding a good web host is understanding your own needs and finding a plan that best matches them for a price you can afford. No one really needs “unlimited” resources, and hosts that offer unlimited storage and transfer know that. They understand that most well-designed websites actually have pretty modest resource requirements. If you know you’ll need a lot of storage space or bandwidth, you’re always better off finding a host with hard limits and paying a little extra.

1. Watched the most awesome machinima ever, Red vs Blue
2. Worked on InkBlot and WebComic

That about covers it. I really need to get out more.

There are, however, two things that Microsoft did correctly with Internet Explorer. The first is conditional comments, which allow knowledgeable web designers and developers to gracefully hide any CSS hacks or special IE code in simple HTML comments that other browsers ignore.

The second, and slightly less renowned, is the IE box model.

Understanding the box model can take a little effort if you aren’t already familiar with web design and development, so let’s use an example. We’re going to have a friend build a box for us, and it needs to be four feet wide exactly. We need exactly one foot of space around whatever it is we’re going to stick inside this box for padding material. And just for the sake of argument, let’s say that the box we’re building has really thick one-inch sides (it’s an awesome box).

Our box-building friend needs certain information from us to build this box, specifically the width, the padding (space between what we’ll be putting in the box and the box edges), and the border (or sides) of the box. We know that our box border is going to be an inch thick, and we want a foot of padding. We also know that the box has to be exactly four feet wide. How wide is the box?

If that looks like a trick question, you’re right. Anyone with any sense would say that the box is four feet wide. If we were to describe our box using CSS, it’d look something like this:

.box {
 border: 1in solid;
 padding: 12in;
 width: 48in;
}

How wide is the box? According to the CSS specification, the box we just created is actually six feet and two inches wide. This is because CSS defines the total width of a box as the width plus the padding plus the border (48in + 12in * 2 + 1in * 2). To get a box that is exactly four feet wide, we actually need to do this:

.box {
 border: 1in solid;
 padding: 12in;
 width: 22in;
}

As you can imagine, this is incredibly counter-intuitive. It also prevents us from doing cool things with relative dimensioning. Let’s try a more abstract example, shall we? We need two boxes this time, and like our last box we want a foot of padding around the contents of our boxes. They also need to have those one inch sides. We’re going to stick them into a storage room, and each box needs to take up exactly half of the width of the storage room. Unfortunately, we have no idea how large the storage room is going to be. All we can tell our box building friend is that each box needs to be half the width of the storage room (or 50%). How wide are our boxes?

Again, this probably looks like a trick question, and it is. Regardless of the size of our storage room, our boxes need to be exactly half as wide as the storage room, or 50% of the storage room width. If we were to describe these boxes using CSS, they might look something like this:

.box1,.box2 {
 border: 1in solid;
 padding: 12in;
 width: 50%;
}

But wait! We already know that CSS determines the total width of a box by adding the width, padding, and border together. That means that these boxes are actually half the width of the storage room plus another two feet and two inches. This is also where the CSS spec fails us, as there’s no way to figure out what the actual “width” of these boxes without knowing the exact width of the storage room.

This is also where the IE box model steps up to the challenge. The IE box model is actually a rendering bug, one that has resulted in at least a few CSS hacks to alleviate. It also makes a lot more sense than the CSS spec, however, because when we tell Internet Explorer that the width of the boxes is fifty percent it assumes we mean the total width. Any additional padding or borders are subtracted from this width, not added to it. With the Internet Explorer box model, we can get exactly the boxes we need, regardless of the size of our storage room.

All modern browsers (including IE8, shockingly) support a “box-sizing” property of some kind that allows us to us the IE box model for creating boxes. Ironically enough, the only browser that doesn’t support “box-sizing” is Internet Explorer 7 and below. And unfortunately, the IE box model bug was corrected in Internet Explorer 6 (previous versions are essentially always stuck in Quirks Mode). So what if we want to use this more flexible, intuitive way of creating boxes in more recent versions of IE?

The answer is Quirks Mode. It’s actually really easy to get Internet Explorer 6 into Quirks Mode; including anything before the DOCTYPE declaration will do it, and the XHTML specification states that an XML prolog should be included prior to the DOCTYPE declaration. By sticking this simple, completely valid line of code into our documents, IE6 switches over to Quirks Mode and we get the desired box model.

Quirks Mode also “corrects” a number of other IE rendering issues. In fact, by forcing Internet Explorer into Quirks Mode, we’re essentially limiting our design and development needs to only two browsers: standards compliant browsers and legacy (Internet Explorer) browsers. Quirks Mode has it’s own set of, uh, quirks, but overall it’s much easier to correct for these than to spend hours implementing various hacks that may or may not break with the next version of Internet Explorer.

Automatic, however, themes the site based on the current season (Autumn, Winter, Spring, or Summer) as well as using special themes for holidays (Christmas, New Year’s, etc.). Most holidays fall on the same date every year, so this is really easy to do. Some are a little trickier, though. Thanksgiving, for example, is the fourth thursday of November, a target that can shift from year to year.

A little searching came up with a simple PHP function that’ll determine when Thanksgiving actually is during any given year using the strtotime() function. We can give it something like “thursday, novemeber 2008 + 3 weeks” and it’ll spit out the appropriate timestamp for Thanksgiving in 2008. Easy.

Easter is a little trickier, though; there’s no simple way to get the “first Sunday after the first full moon on or after March 21.” Just reading that makes it sound like it would take some kind of archaic, trance-inducing mysticism to figure out (assuming you don’t have a calendar handy). What’s one to do? Why, use the easter_date() function of course!

easter_date — Get Unix timestamp for midnight on Easter of a given year

Just give it a year and this bad boy spits out Easter like there’s no tomorrow. Or something. It’s not perfect but it gets the job done well enough, and I nearly died of laughter when I saw it. I just love the fact that there’s a dedicatd PHP function for computing what day Easter falls on in a given year.

Still, Chrome is very polished, and after just a couple days of use I don’t see any reason to return to Firefox. Part of the problem (with me and Firefox, that is) is that I’ve never been huge on add-ons. I don’t care that I can use Firefox as a feed aggregator, MP3 player, instant messenger, etc, etc. I get it, I really do, and I just don’t care. Chrome gives me exactly what I want a super-fast, elegant web browser that lets me interact with the web.

The only add-ons I actually used were Adblock Plus, Delicious Bookmarks, Firebug, and Google Notebook. I imagine Chrome will get some kind of integration with Google’s web services in the future, Notebook included. I use Delicious largely because of the Firefox add-on, and Google Bookmarks integration with Chrome seems like a no brainer that will show up sometime in the future. That leaves Adblock Plus and Firebug, both of which I really can live without. Chrome already looks to include some interesting developer features; it’s doubtful they’re up to Firebug levels yet, but as with many Chrome features there’s room for improvement (Chrome is only at version 0.2 here, folks).

That leaves Adblock Plus, and while I love it, my ad blindness has gotten to the point that I glaze over pretty much anything that looks even remotely like an add unless it blares sound at me. In that case, I immediately close the page and never return to the offending site.

For what I use a web browser for—that’d be browsing the web—Chrome is already a better choice than Firefox. And if (when?) Google gets around to integrating their various web services into the browser it’ll be a truly killer application.

Update: This article over at cnet is interesting, not least of which because if you read the list closely enough—or even just sort of glance at each point—it sounds very much like the author wants Chrome to be Firefox. The only point I even kind of agree with is #4: Saved Sessions. Chrome hasn’t crashed on me yet but there always the chance it will (especially this early in the game), and with my luck it’ll happen when I have thirty mission-critical tabs open. As for the other points…

1. I’m not totally sold on roaming profiles, which is to say that I don’t see why a web browser should be storing a significant amount of data on a users machine anyways. As the author mentions, Google already has an extensive search history service; extending to a web history service that Chrome links into would be great.
2. Continuing from point one, I don’t see why I should have to deal with managing local bookmarks anyways. Chrome should be using Google Bookmarks, which solves problems one (for bookmarks, anyways) and two. I’ve been trying to get away from using local bookmarks for quite awhile now.
3. I think I covered plug-ins (sic) pretty well above. They’re coming anyways, so there’s no need to gripe.
4. As mentioned, I agree with the first half of this. I couldn’t care less about the second half. I’m really glad Google did away with that obnoxious warning.
5. I’m apathetic to full-screen mode. I can’t remember a time I’ve ever really used it, but it wouldn’t hurt to add it in for those that need it.
6. As much as I love customization in some cases, I hate this idea. I don’t see it happening anyways, though. One of Google’s proposed goals with Chrome is to make the browser as unobtrusive as possible, which is counter to the idea of tricking out the browser window with custom themes.
7. I think this shows a distinct misunderstanding of the Chrome application paradigm. There’s no way to drag an application into a Chrome window because applications aren’t tabs. The whole idea of turning a website into a Chrome app is to do away with what little browser interface there is and focus entirely on the website. Being able to drop an application into a Chrome window doesn’t make sense because the two are completely different things.
8. Like plug-ins, this, too, is coming.
9. I don’t quite understand this… maybe I haven’t encountered enough pop-ups in Chrome, but from what I’ve seen I actually prefer the way it handles them.
10. And finally, this is the worst idea ever. The omnibox is a spectacular idea, one that other browsers would do well to mimic. Having used it I’m surprised no one else thought of it. The fact that it automatically adds searches from any site you search on is a major win, as well. The omnibox gets my vote as the most elegant, efficient, and effective address/search bar yet conceived.

Unless you’ve somehow completely avoided tech news for the last two days you’ve no doubt heard of Google Chrome, the long-rumored, finaly-realized Google web browser. I’ll admit: I’m both very excited and a just a little worried about this. I’m excited because, well, it’s a Google web browser. I’ve been waiting for a Google web browser since it was just a rumor long ago. And with the plethora of Google services now available, a web browser seems a natural next step for the company that offers everything else.

It scares me (just a little) because, well, Google is everywhere now. They’re ads are found pretty much everywhere. I’m using Google Chrome right now to write this post. Gmail manages all of my e-mail accounts. Google Calendar keeps my agenda in order. Google Reader keeps me up to date on the plethora of sites I pull news from. Google Talk keeps me in touch with friends (well, sort of). Youtube provides me with a venue to inflice my full-motion presence upon an unsuspecting web.

It’s just a little disturbing when you realize you relay on a single source for nearly everything. I haven’t gone into full tinhat mode yet, though. I don’t have to use Google’s products (even if they are often the best). And it’s not like I’d be totally lost if Google vanished from the Internet. There are plenty of other services available I could use.

Anyways, some things I like about Chrome so far:

• It’s fast. Really, really fast.
• I love the interface. The screenshots that are available make it look kind of junky, but it actually integrates really nicely into Vista. It’s also very minimal and out of the way, something I’m a big fan of as far as interfaces go.
• The Omnibox is spectacular. It’s like the Firefox 3 Awesome Bar with fully integrated search. I’m still a little confused as to how it adds site search engines… for awhile it seemed like it wasn’t adding any, but now it seems to add them after only one search. All you have to do is start typing the name of the site (Amazon, for example), hit tab, and search away.
• The new tab page is a nice touch. It integrates speed dial, search, and recent bookmarks/closed tabs into one convinient location.

And some missing features I’d like to see:

• Integrated spell check.
• Full-page zoom.
• Amazingly, there doesn’t seem to be any way to force mailto: links to open in Gmail, which is something that’s so painfully obvious I fully expect to see it in a future update.

Finally, there’s one thing that really bugs me about the Google Chrome coverage. I keep seeing comments about how “designers/developers are going to have to design/develop for yet another browser *roll eyes/sarcasm/spite/whatever*”, most notably in this cnet article. I’m really tired of seeing this. It’s the underlying rendering engine that matters, not the browser itself, and Chrome uses Webkit, the same rendering engine as Safari. While there might be some tweak or something going on under the hood (particularly V8, in this case), it’s still largely Webkit, which means that anything that works in Safari will probably work in Chrome.

And to be honest, Gecko (Firefox), Presto (Opera), and Webkit (Safari and Chrome) are so close in rendering quality (per web standards) that it’s often safe to assume that a well-designed page will render the same (or almost the same) in all three if it looks right in one. A good designer won’t bet on that, of course; it’s always best to test. The only true, outstanding annoyance is Internet Explorer, however.

And with any luck, Chrome will destroy it.

Update: The problem with upgrading is that various aspects of the site require alterations to core WordPress functions. I just spent the last couple of hours trying to figure out how to make what changes I need via plugins and I’ll be damned if I can figure it out.

Part of the problem (I think) is that the changes I’m making require altering functions that the WordPress API currently has no hooks for, meaning that the only way to alter htem is to do so directly. Which is bad, ’cause that means everytime I upgrade WordPress I have to go back and make the same changes all over again.

Prior to WordPress 2.6 I was doing the hacks directly in the core files, but now that I’ve upgraded I’ve taken the marginally better step of placing all of the hacks in the functions.php file that accompanies the site theme. It’s not perfect; I still have to modify the core files, but now I’m simply commenting out the functions I need to modify, copying them into functions.php, and making the necessary changes there.

If anyone reading this knows a lot about WordPress: is there some easier way to modify the way WordPress outputs category, page, and link lists? The functions that handle the building of these lists seem to be pretty low-level and inaccessible via API hook. I’m probably (hopefully) missing something, but I haven’t seen anything in the documentation that shows how I would get to these functions.

I also didn’t get quite as much sleep as I’d hoped to yesterday, but I’ll do my best to get the comic up tonight.