Let there be Dust-Me Selectors. It’s is a Firefox extension that finds unused CSS selectors and notifies you so you can remove them from your stylesheets. Awesome.

Post from: Capsize Designs

Eliminate CSS Framework Bloat with “Dust-Me Selectors”

Let me know if you have! You can comment here, email me from the contact page, or post it on the BlueTrip forum.

Post from: Capsize Designs

Used BlueTrip For A Site?

Firefox, on one hand, is a whole different ballgame. It’s huge, has a hugely popular plugin architecture, open source, etc. It’s the big papa of the standards-compliant crew, in direct opposition with IE.

But I’d say Opera and Chrome are pretty closely related to each other. They’re both fast and nimble, look pretty good, have the same “speed dial” type of start page, have a “omnibar” more or less, etc. Why is Chrome taking off while Opera sits around on its hands is a mystery to me. Unlike Chrome, Opera is reliable (read: NOT in Beta), has support for themes and widgets, can run on ANYTHING (Linux, BSD, Solaris, and many mobile phones with Opera Mini, to mention just a few), and 9.6’s default theme looks really slick.

So why this unfair market percentage? Why doesn’t Opera ever get the respect it deserves? Is it because of Chrome’s overly hyped V8 javascript engine? Or do people favor WebKit that much? Or just the fact that it’s Google?

Post from: Capsize Designs

What Does Chrome Have That Opera Doesn’t?

Come on, people! While, the Linux package msttcorefonts installs a lot of common Windows fonts on a Linux distro, it can't be assumed that all Linux users go to the trouble. In fact, more than 1/3 of them don't. Please, please, please, throw in a common Linux font before your generic fallback to catch these guys.

Linux users may only be a small percentage of the overall web population, but we're a very important one. We work the servers. We hack the CLI. We are some of the most influential players in the FOSS movement. And we're good at hacking (but nowhere near the BSD guys...one of them hacked my dog and turned it into a cat right in front of me). So please, give us a little respect. I think we've earned it.

Here are some sample font-families for you (taken from the Christian Montoya article on Vista fonts with added support for Linux):

Any feedback? Fonts that I missed?

Post from: Capsize Designs

Linux Users Need Font Families Too!

However, I've stuck to my guns and followed the wisdom of Eric Meyer, among others. The only true argument I can see against a reset stylesheet is that they add unnecessary code, which usually amounts to less than 20 lines, sometimes only taking a couple. With most sites using over a thousand lines of CSS, I think we can make do with the extra dozen.

In my defense of reset stylesheets, I'm always running across the argument that the * selector does everything you need, using something like:

Rather than manually listing each item you want to reset, as is the case in Meyer's reset and many others, this method resets everything. This brings up the obvious counter argument: you're resetting things you don't want to be resetting. I usually quote Christian Montoya's argument that this puts a strain on the rendering agent and resets things you don't really want to be resetting.

However, in all of my searching, I've never been able to find any kind of benchmark or study that proves that the * selector really does hurt performance. The pride in me wants to believe that the * selector really is a bad practice, but logically it seems that it would only really apply to elements on the page. So say for example you have a page with only a form, then the * selector would only apply to elements of the form, such as the inputs, the fieldset, the legend, etc. That doesn't seem unreasonable. Then say you add a header, then you've got it affecting the unordered list of nav links, the links themselves, etc. So what? At what point does this become a nuisance? And how much is there really that it could possibly be resetting that Meyer's reset doesn't already include?

With the * reset being so easy to remember and it being widely used in spite of being deemed a bad practice, I'm tempted to take it up, and I hope some of you guys have some insight here. If this really is a bad practice, tell me why. And if not, please let me know so I can stop arguing about it!

Post from: Capsize Designs

Is The * Selector Really That Bad?

But I have spent some time with JSF. JSF (JavaServer Faces) is a JSP framework (if you can call it that), which aims to separate logic from presentation and simplify the development process. And I'm here to tell you first hand that JSF is a big old mess of "don't go there." Here's why:

Relearn HTML - The Crappy Way

JSF replaces a good chunk of basic HTML tags with their custom ones, because that way you can insert dynamic content into them. Unlike PHP, where you can just echo anything you want in the value of a button, for example, using something like this:

In JSF, that would become this:

Why, Java? Why!? And this isn's just for buttons. It's also for any form input, the form itself, any outputted text, links (anchors), images, etc.

Don't Touch My Attributes!

JSF also has a tendency to screw around with your attributes which makes CSS, and sometimes even HTML validation, pretty difficult. For example, in a form, if you do something like the following:

...instead of outputting input type="text id="nameinput" you get input type="text" id="myform:nameinput". See the problem here? Try styling myform:nameinput with CSS. It thinks nameinput is a pseudoclass, and of course it isn't, and it screws everything all to hell. Good luck validating with a fake pseudoclass! Yayyy!

Hope You Don't Need Logic In Your Views!

JSF doesn't include any way to put if or foreach or anything like that. Take a moment to let that sink in. By default, you CANNOT use any logic code in a JSF page. So say you need to loop through all the products that match the user's search query. Can't do it. Nope. Nuh-huh. This is ridiculous.

That means that you have to resort to using JSTL based tags, which enables you to use and in your views. The problem is that JSTL and JSF don't really mesh well, which can start throwing some crazy errors. Boo, Sun Microsystems. Boo.

XML Based Navigation Rules are BAD!

Say you are using a login form, and you want to send a user to one page if the login is successful and another page if it's not. You would do this using this beast of an XML file known as faces-config.xml. You'd need to set up a navigation rule in there which only activates if the user is coming from a certain view, and decides where the user goes based on what the login function returns (usually a string of some type such as "login_successful" or "login_failed").

This works, don't get me wrong. Unlike the rest of JSF, this is doable. But why go to the trouble? Why can't we just put redirects in the code itself? Why the need to add XML into the mix? It's just another layer of confusion that we could do without.

GET and POST Might Kill You

Getting something from the GET array looks like this:

Any questions?

And that's why I will never be going anywhere near JSF in the near future.

Post from: Capsize Designs

The Trouble With JSF (JavaServer Faces)

My only requirements: PHP and MySQL, really simple, outputs valid code, and easy to theme. So far, the only options I've found are punBB, flexBB, and SMF, though I'm sure there are others. However, I happen to know that anybody who reads this blog must be brilliant, so anybody care to drop some knowledge?

Post from: Capsize Designs

The Search For A Standalone Forum

Free and Open Source (GPL, anyone?)

Heckle if you wish, but this is a biggie for me. I need to be able to download and play with it for free! I need to be able to use it on my auntie Barb's beauty parlor site that she's not paying me a dime for. I need to be able to screw with the code and reuse the code and branch off from the code without getting sued.

If you're up on the CMS world, you'll know that ExpressionEngine is one of the main violators here. I won't say that EllisLab needs to Open Source it, because they're obviously doing fine with their $100-$250 licenses, and it's money from that that enables them to support my beloved CodeIgniter. But it's frustrating to me that for any client that's not willing to fork over the license fee, I'm stuck with either a stripped down version of a good CMS or something else that's not as fit for the job. So I'll just whine about it for a minute and move along.

Dead Simple Templating System

We want a CMS that you can fully build a site on without having to code ANYTHING server side (this means XHTML/CSS/JS only). Ideally, there is a web interface where you can edit templates, but these templates can ALSO be edited in a regular old text editor (take note, Textpattern!). Pages should be nothing more than HTML with little tags thrown in (note: tags does NOT mean PHP variables! That means you WordPress, SilverStripe, etc.) for the dynamic content.

I have yet to find a winner here. Textpattern gets close, but the inability to edit in a text editor just sucks, and the use of the word "Forms" instead of Snippets or something like that is strange. ExpressionEngine also comes close, but the tags are a little annoying. Some need an "exp" prefix and some don't. Some auto-close, and some don't. I like the XML style greater than and less than symbols too, over the curly brackets. But now I'm just whining again.

Easily Distributable and Interchangeable Themes

Pretty much all of the front runners are going to have at least a few themes you can use, but many of them don't have an easy way to install them or change them. For example, changing themes in Textpattern means a long process of copying and pasting into the web interface template editor. Same with ExpressionEngine. Drupal is also a little annoying here.

WordPress gets this one right. Just have a themes folder, throw a theme in there, and change it with one click in the admin interface. Although I'm not the type to yank somebody else's theme for my site, I do like to start with a theme that has the basic layout I'm looking for, and work from there. I find this is a good way to stay up on how others are achieving a certain layout, and brush up on messing with others' code. Plus, it saves some time over starting from scratch.

Officially Supported Modules, and LOTS OF 'EM!

We all love a fully featured CMS, but we've also all pulled our hair out trying to get those hacked together "features" to work. Take Drupal, for example. Say you want to throw in a webshop. You'll probably go hunting through the dozens of eCommerce plugins looking for the perfect one, and try out a few only to find that they're no longer supported or they don't mesh with the current version of Drupal or WHATEVER. It's annoying. Textpattern's plugin community is worse. Don't get me started.

There are of course some good things that come out of having a user-contributed plugin system. You get a lot more than you would if only the CMS's own developers were contributing, for one. But the quality over quantity argument never holds very strong. We need a CMS that's fully featured by itself! ExpressionEngine wins here. It's list of features is pretty crazy, and all of these modules are developed by EllisLab, not some guy in the Netherlands (no offense to any readers from the Netherlands! But you're so far away!) SilverStripe tries, but it's still pretty immature and many of the modules are very early in development.

PHP and MySQL Supported

This may not appeal to all, but I think in order to get the community a good CMS needs, it needs to serve the broadest base, which is PHP and MySQL. I'd also like to see support for SQLite, but I won't get greedy.

Ridiculously Simple Admin Interface

Let's face it. Our clients are idiots. They need idiot-proof GUI's. And 90% of CMS's out there are by no means idiot-proof. I, a web developer, was even a little intimidated when I saw Drupal's interface. WordPress's is better but not great. SilverStripe's seems easy, but it's kind of too easy. Editing pages is easy, but anything other than that is kind of a mystery. Say you want to change basic site config options. There's no "Settings" tab. Huh?

Textpattern wins in my mind. It's dead simple. It assumes (like WordPress) that the first thing you want to do is add content, so it takes you to that page when you log in. It hides all the complicated options away where the client can't find them unless he or she tries. It just makes sense.

Why It's Impossible

So we've listed what we want: Free, open source, easy templating, easy theme distribution, official and numerous modules, and an easy to use admin interface. This is one big stinking contradiction.

First of all, the free and open source part seems to contradict the official modules part. The whole point of open source is user contribution, and you're not going to see many open source projects that rely only on core developers' work. It's stupid. You might as well just close it up and copyright it.

And the templating engine I have imagined doesn't mesh well with distributable themes. An easy to use, flexible, templating interface usually means that the templates are stored in the database, which makes it pretty durn difficult to edit them in a text editor, and even harder to zip them up and distribute them to other installs, and even harder still to install them on other installs. So it looks like you're going to either have one or the other.

Finally, the feature rich CMS I'm looking for (complete with an official forum module, gallery module, registration and profile module, etc.) has so many things to do that it's near impossible to make it easy to use. You can only dumb things down so much before they get too simple, as, in my hopefully humble opinion, is the case in SilverStripe.

Bottom line: we're going to have to pick and choose which features are most important to us on a case by case basis, and pick the CMS that is closest to that feature set. Looks like, in this case, nothing is perfect.

Post from: Capsize Designs

The Perfect CMS (Or Lack Thereof)

For anyone that uses Pidgin, Gaim, Miranda, etc. on multiple computers, this could be a godsend. It's just a pain to maintain a few different sets of usernames and passwords/protocols/options on different computers.

9. Route faxes into it.

A few inventive users have managed to have all company faxes thrown into a shared "Faxes" folder which anyone in the company could access. This way, you're not wasting paper, you're not wasting time emailing faxes around, you don't have to worry about messing with networking the faxes all over the company network, etc.

8. Keep a toolbox of portable programs.

PortableApps and portablefreeware.com both make a pretty good presence on the net by distributing free portable (meaning they don't have any "install" process and they leave no traces on the computer because they're usually run from a flash drive) programs. Now, instead of keeping them on a flash drive, you can throw them into Dropbox.

7. Manager source code.

The Dropbox folder can be installed anywhere (i.e., even in /var/www for apache servers), meaning you can run your localhost sites out of Dropbox. This way, you have a (really) simple revision control system and online backups to prevent oopsies.

6. Secure your computer.

One of the slickest Dropbox methods I've seen involves installing a keylogger which dumps all the output into the Dropbox folder. This way, if your computer ends up stolen, you have all of the thief's keystrokes (often names, addresses, etc.) backed up on the web. Another user went above and beyond and installed a webcam which will automatically dump videos into Dropbox for security reasons.

5. Keep a library of desktop wallpapers.

I'm a wallpaper fanatic, and I'm constantly having to transfer from comp to comp on a flash drive so I have every wallpaper on every computer. That's done!

4. Collaborate on documents.

While Dropbox isn't really made for this, it does a pretty good job of allowing people to collaborate documents (reports, homework, source code). Just share the folder with a few people, and you're good to go. You run into a little hiccup with conflicts, in which case Dropbox keeps the one that was finished downloading to the server first, and names the other one something like "Mike's conflicted file 1-1-08". It ain't Google docs where everything is realtime, but it'll do.

3. Keep track of browser bookmarks.

My Dropbox is my sole repository for Firefox (and Opera) bookmarks and other settings I'd like to maintain from computer to computer. It's a nice alternative to del.icio.us.

2. Keep track of design work.

Dropbox is a designer's dream. When you're working on a new site in Photoshop, or making some crazy icon set in Illustrator, just keep the files in Dropbox. That way, if you screw it all up, and you don't have the undo history, you can fall back on the revision history. Also, since Dropbox only sends the parts of the file that are different each time it's saved, saving your 100mb Photoshop file only takes a couple seconds because all it has to do is save the changes. Plus, you have online backups in the case of a hard drive crash, and you can easily ship them mockups out to clients via the public folder.

1. Use it for everything!

Just about my entire /home (or My Documents for Windows users) folder is inside of Dropbox. I keep designs, code, reports, invoices, ebooks, music, photography, whatever. That's the beauty of Dropbox. Just drop it in.

Post from: Capsize Designs

Top 10 Ways to Use Dropbox

First off, I think we all need to realize that this is a VERY early release of a huge project for Google. I have no doubts that it will receive tons of critics based on bugs, missing features, and generally things that Google will obviously be fixing as the project matures. The last thing I want to do is point out the obvious shortcomings (limited bookmark management, few options, no progress bar for page loads, etc...guess I pointed them out anyway), because these things will be coming soon. Google's, while perhaps a bit optimistic and overconfident, is not stupid by any means. So whatever you say about Chrome, please don't go and complain about the lack of things that are almost certainly on the way. Or else you'll annoy me. That's right!

As far as I'm concerned, there are two huge things that Chrome does right on the back end. The first is V8, the Javascript engine that benchmarks prove is easily the fasted engine out there. Another set of benchmarks puts it at twice as fast as FF3 and the latest Safari. While the numbers obviously depend on the system, it seems pretty obvious that V8 (which, by the way, was deemed by Google to be worthy of separating completely and therefore outsourced to a company in Denmark) beats any other Javascript engine out there at the moment, even Firefox's bleeding edge TraceMonkey. And while we all know that benchmarks may not necessarily carry over into real life situations (even well supported projects like SunSpider), I'm here to tell you that I (assuming that I wasn't imagining anything) could tell a noticeable difference when running JS intensive web apps (Gmail, Zoho, etc.)

The other huge back end aspect is the fact that Chrome treats each tab like it's own mini browser. We've all experienced one tab bringing them all down (usually due to a flash hangup in my experience) and Chrome intends to put a stop to that. One tab going down will only bring that one task down, and Chrome's integrated task manager makes it easy to identify which tab is causing the problem and shut it down. Not bad. And I'm here to tell you that, even at this early stage, this feature works, and I'm loving it.

On the front end, there are a few obvious things. The tabs are above the navigation bar instead of below, through I'm hard pressed to see this as anything more than a style choice. It makes sense, given that each tab is to act as it's own process, that each tab should contain it's own navigation bar, and I guess that's the point. But really, it's not as big of a deal as it's being made out to be. So the tabs are an inch higher. Whoopee. And then there's the all-in-one address bar/search bar/bookmark bar. This is also a little too hyped. Modern web browsers (Opera 9.5, FF3, IE8b) all allow you to search the web, your bookmarks, and type in URL's directly into the bar. As far as I can tell, the only difference is that they stole my beloved dedicated search box in favor of simplicity.

And there are a few pretty major things that I'm under-excited about. One is the addition of Google Gears. And along with this is the ability to launch web apps in a dedicated toolbar-less window, which treats it much like a desktop app. I guess that's cool and all. But I feel sure that I'll never use it. I have no use for using web apps offline (simply because I'm never stuck without web access) and I don't like the idea of trying to trick myself into thinking I'm using a desktop app (why take away the navigation bar? It's the internet. Let me treat it like the internet). But I don't want to downplay these features because I can see how they'd be useful to a certain type of user.

One thing that I'm really disappointed about so far is the inability to use Chrome with Google services. I want Chrome to sync my bookmarks with Google bookmarks, and the built in ability to take some notes with Google Notebook, and automatic syncing with Google Web History, etc. As far as I can tell, there's no way to tell Chrome about your Google account because there's no use. I'm not sure if this is in the upcoming featues (perhaps one of the readers can let me know), but I sincerely hope it is, because this seems like one of the great possibilities of having a browser made by Google. Perhaps Google would rather leave this out in favor of neutrality, but I'm pretty sure Google knows better than that. Give me the option!

So here's my main viewpoint, for anyone who sat through all that. One of the things I think Google has always been good at is getting the IT world talking. They've done it once again. As far as I'm concerned, nobody is doing more to further web application development and the internet world in general than Google, because nobody consistently comes up with new ways of doing things like Google does. I won't say that Chrome will be a huge hit. With a top notch product like Firefox only getting around 19% of the market share, and IE 6 still getting a huge chunk (not to mention the big ugly beast IE7), there is almost no hope that the better product will get the market. What I am saying is that Google is good for competition. V8 will force competing browsers to revolutionize. The task manager will get developers thinking. Even the upper tabs, while maybe being blown up, will get browser developers thinking outside the box. While I'll still probably stick to a mix of FF3 and Opera, I'm excited to see what the future brings for Chrome, and what the future of Chrome may do to the future of other browsers.

Post from: Capsize Designs

My Humble Impressions of Google Chrome