Skip to: site navigation/presentation
Skip to: Thoughts From Eric

Archive: 'Tools' Category

Wanted: Excerpt Exacter

So after I once again published a post without filling in the excerpt, thus forcing me to go back to fill it in later, I tweeted in a fit of pique:

I need a WordPress plugin that won’t let me publish a post until I’ve filled in the excerpt field. Anyone got one?

To which I got a whole lot of responses saying, in effect, “Oooo!  Good idea!  I need that too!  Let me know when you find one!”  Some of them came from people running fairly high-profile blogs.  The need clearly exists.  A couple of responses were of the “I could do that!” variety, so I thought I’d post here so as to describe how I think it ought to work from the user’s perspective, and then we can hash things out in comments and someone can code it up and make everyone happy.

So really what I want is, when I push the “Publish” button in WordPress, the plugin checks to see if there’s an excerpt.  If not, one of two things happens:

  1. The plugin throws up a warning dialog telling me that I’m about to post without an excerpt (again).  If I say “Okay”, it goes ahead with publishing.  If I say “Cancel”, it returns me right back to where I was, which is the “Write Post” page, with all the data intact and unaltered.

  2. The plugin returns me to the “Write Post” page with all data intact and unaltered, and puts an error box at the top of the page telling me I forgot to write an excerpt (again) and that it won’t let me publish until I fix the problem.

One or the other.  I think I like #1 a little better, but I’d be good either way.  I’m open to other approaches as well, but I don’t think the plugin should rely on JavaScript, as that means leaving out people who don’t enable JavaScript or post from JS-incapable devices.

I would do this myself, but I’m a little wary of the “return to the page with all data intact and unaltered” bit, which I would imagine is pretty easy to mess up.  Thus I’m putting it up here as a semi-Lazyweb post so that someone else, someone with more experience with WordPress and plugin authoring, can do it right and quickly.

Okay, who’s on it?

WordPress Adminimize and Latest Tweet Plugins

Just because the world hasn’t suffered enough bad PHP coding, I released two WordPress plugins today.  One pares down some of the admin UI, and the other is a Twitter plugin.  Because the world doesn’t have enough of those, either.  You can find links to them on my WordPress Plugins and Hacks (but I repeat myself) page, or if you can go straight to their respective pages, which I’ll link here in just a second.

Okay, second’s up!

MW Adminimize (currently v0.5)

This plugin pares down the page-topping links interface that shows up on every WordPress administration page.  I got tired of seeing my blog name in Big Huge Letters all the time, and I kept forgetting that the “Dashboard” link was hidden up in that black bar.  So I fiddled with some arrays and wrote some CSS, and got something that chews up less screen real estate while also being easier for me to use.  Maybe you’ll feel the same, in which case, enjoy the plugin.

In addition to that, I pared down some of the elements of the “Write Post” page so that they took up a bit less space, and compressed the stuff around the post title’s input field.  Those styles are marginally less stable and may cause weirdness for you.  If so, sorry.  Let me know what goes south and I’ll try to get it fixed in a future version.

Oh, and I also included a CSS rule that makes the background shading of moderated comments more obvious, though not in my usual eye-scarring fashion.  It’s just a little darker and notably redder.  That doesn’t have much of anything to do with minimizing the admin UI, but I thought it was a huge improvement.  I might take it out in a future version if the crowd shouts otherwise.

MW Latest Tweet (currently v1.0)

This one does what you’d probably guess: shows the latest tweet.  Why not just use something like Twitter Tools?  Actually, I tweeted about that.  A further survey of the options available turned up nothing that was quite what I wanted.  So of course I started hacking away.

The basic mechanism is that any time a page that calls this plugin is loaded, it either returns the last-cached tweet or asks Twitter for the latest tweets so it can cache the latest of them.  If Twitter fails to respond (I know, what are the odds?), you still get the cached tweet.  This is similar to SimpleTwitter, from which I took much inspiration, but is more advanced in a number of ways.  Among the options I wanted to have and thus created was having “human time” intervals as an option, and also the ability to filter out @replies (because I personally wanted to display my latest broadcast tweet, not my latest reply).  While I was at it, I decided I’d throw in the ability to define your own XHTML output, so that if the plugin failed there wouldn’t be partial or mangled markup shoved into your pages.  Also, since I cache the whole of the latest tweet, you have all kinds of stuff you can output should you so desire.  It’s all documented on the plugin’s page.

This is not a widget, because I’ve never written one and didn’t have time to learn (due to all the time I was burning just trying to write the plugin).  You have to edit your template/web page PHP in order to call it.  That’s just the kind of DIY guy I fly.  Er, I am.

Also, I’m a wee tiny bit proud of how I styled the labels on the plugin’s Settings page.  I’ll probably write that up in a separate post.

That’s about it.  If anything goes wrong or is unclear, please do let me know in the comments.  Share and enjoy!

Expressive Sculptor

For those of you using Microsoft Expression Web, a free pre-release trial version of CSS Sculptor for Expression Web was announced by the WebAssist folks on Wednesday at MIX08.  So now you don’t have to put up with those snooty Dreamweaver users throwing you the mëtäl hörns every chance they get—throw ‘em right back!  Røck!

If you’re curious about CSS Sculptor, I posted in some detail about it when it was first released in August 2007, and there’s of course plenty of enthusiastic copy about it on the WebAssist site.

One thing that’s different about the Expression version as compared to the Dreamweaver version is that it doesn’t have an “Apply” button to apply the input CSS ito the preview window.  Instead, changes are instantly reflected in the little preview.  It’ll be interesting to see how users react to that, since it could mean that the previewed design shatters as the CSS is updated, and then snaps back together upon further changes.  Is that good or bad tool usability?  Hard to say; it could scare people into undoing the shatter-change and never pushing forward, but it could also help users more quickly gain a deeper understanding of CSS by seeing how things come apart and then go back together.  I guess we’ll find out!

Manhattan Problem

It’s not every day I uncover a case involving the botched theft of information about nuclear weapons.

Here’s how it went down: in the infosthetics feed was an entry about a video regarding nuclear stockpiles around the world and the effects of a nuclear explosion in New York City.  The video was produced by Chimp on a Chain for Good Magazine.

That’s a long-standing area of interest for me, so I watched it.  When I got to the New York City portion, something started to bother me beyond the obvious horror of the scenario.  The point of detonation, the explosive yield, the elapsed-time intervals, the radius distances—all seemed very familiar, like I’d seen them somewhere before.  And I had.

They were nearly all taken verbatim from the New York City scenario found at the Atomic Archive.  I could find only two differences.  The first is that the total death toll given in the video is slightly higher than that in the Atomic Archive’s scenario.  Otherwise, all the numbers matched up.

The second difference is really a major error on the part of the video’s makers: they dramatically under-represent the areas of damage.  For example, the ten-second ring’s (found at 2:33 in the movie) radius is labeled with the correct distance (2.5 miles) but the circle placed on the map is much, much too small to be 2.5 miles in radius.  The circle doesn’t even cover the breadth of Manhattan Island, whereas an accurate plot would have it stretch across the Hudson River on both sides into New Jersey and Long Island.  You can see this in part 5 of the Atomic Archive’s scenario, or on a HYDEsim plot of the same scenario.

The video seriously misrepresents the area of damage that would result from such an incident, making it appear much smaller than it would be, and I just can’t fathom how or why they would get that so wrong.  Even assuming they mixed up the meanings of “radius” and “diameter” doesn’t appear to explain it.  The ring distances shown correspond to a three-kiloton explosion at most, not to 150KT.

That’s the botched part.  So where’s the theft?  There is no credit whatsoever given in the video for the material’s source.  There is a reference to the Archive on the video’s page at Good in the “Resources” box, but the material in the video has been used without permission—I checked this with the custodian of the Archive—as required by the site’s policy.  Even if one could argue this is a case of not needing permission on non-profit grounds, attribution is still required.

It would almost be worth subscribing to Good so that 100% of my payment could go to the non-profit of my choice, as the site promises, except I’m limited to their choices of non-profits and none of them appear to be charged with educating magazine publishers or video artists about the niceties of copyright law, intellectual property rights, or even just plain common courtesy.

CSS Tools: Reset and Diagnostics

I’ve hinted and teased and promised, and I’ve yet to make good on any of it.  I’m sorry.  Can I make it up to you?

Okay, then, here you go: a permanent home for my reset styles.  It takes up residence in a new “CSS” subsection of the Toolbox section of the site, along with my efforts to create a generic set of diagnostic styles.  In the case of the resets, I’ll increment the version number and date whenever I make changes, and probably maintain an archive of previous versions.  Not that I expect that to happen with any frequency, but you never know.

The diagnostics are a lot more experimental and thus aren’t so formal as to have things like version numbers.  If I change anything of specific interest there, I’ll try to write a post about it.  I should get around to writing about the shown-in-page stylesheet one of these days, though I figure anyone can view source and work out the particulars without too much effort.

In Search of Q

In an effort to get a handle on my taskflow, I went looking for an organizer application.  So far as I can tell, what I want doesn’t exist, but maybe someone can point me to it.

What I really want is a push queue for documents and other data fragments.  I’ll call it “Q”, both for the obvious phonic match as well as to score a little ST:TNG joke plus make a Cleveland arena reference.  The latter two work because I sort of envision the application as being a very powerful being as well as a large gathering place for data.

The way I envision it, I drag a file onto the main Q window and it’s added to the general pool.  Every item in Q can be labeled, tagged, commented, and otherwise meta’d half to death.  The queue can be sorted or filtered on any number of things—file creation or modification date, Q addition date, file name, containing folder, tags, labels, and so forth.  Also, every item can be assigned a due date.

When I double-click on anything in Q, it opens the original file just as if I’d double-clicked its Finder icon.  (I’m an OS X user, but translate “Finder icon” to whatever the equivalent words are in your OS of choice.)  So really, Q is maintaining a pool of aliases to the original files, plus any associated metadata.  In that sense, it’s like iTunes set to not copy added music to the iTunes Music folder in your home directory.  Yes, some people run it that way.  And like iTunes, the ability to create smart lists based on tags and comments and such would be really awesome.

I’d find Q deeply useful because as new tasks come in/up, I could drag in whatever file(s) relate to those tasks so that I don’t lose track of what I have to do, quickly tag them and set a due date, and continue with whatever I was working on.  There’s room for tons of even more useful features like synchronization across multiple computers, the ability to accept any fragment of data at all as opposed to files, and more, but the core need is a task queue.

To illustrate this with some examples from my recent workflow, I would drag in a copy or two of the IRS W-9 form, a couple of e-mail messages, an invoice, and a Word document containing a set of interview questions.  The W-9s would get tagged by the clients’ names, the invoice would be tagged and flagged, and so on.  The real key here is that they’d be add-sorted by default, so I can work on them first-come-first-served.  Of course, other approaches would be possible with other sorts and filtering.

It seems like, with all the GTD mania floating around, someone would have come up with this solution already, but my searches have so far been fruitless.  I tried a couple of applications that seemed like they might be close to what I want, but they weren’t.  Am I just using the wrong search terms, or is this something that just doesn’t exist yet?

Another Soul Lost to SketchUp

Hitting a link shared by Unstoppabot, who really needs to get around to fixing his feed linking policy (“View this site”?  Lame!), I was seized with another spasm of appreciation for the deliciousness that is SketchUp.

I did a moderate amount of 3D modeling back in the day.  The specific day in question would be the one where we all thought that images of rendered 3D models and, whenever possible, blobs of text were the absolute last word in Great Web Design.  Remember that?  Wasn’t it fantastic?  When every page title could be a bunch of extruded and beveled sans-serif letters viewed slightly from above, with the whole mess of text angled away from the observer?

Good times.

So anyway, while I was cranking out renderings of page title text and university logos, I also spent some time creating scenes of other stuff.  You can find some of the results if you dig deeply enough here on meyerweb, but that’s not my point.  What I’m trying to say is that I enjoy a bit o’ three dee more than most, and have some knowledge of how difficult it can be to construct models.

When I first heard about SketchUp, I was intrigued but didn’t really buy into all the hype.  It couldn’t be that easy, could it?  And then I watched someone using it—at An Event Apart, as it happens; and no, it wasn’t one of the attendees—and was captivated.  I downloaded the installer while I was sitting there, watching him create and modify shapes as easily as sketching them on paper.  And then I left it uninstalled, because I was afraid of what it would do to my free time.

A few days ago, I finally broke down.  I actually did have a legitimate reason to install and use it, a really good one, but of course I’d been waiting for any reasonable pretense to launch the .dmg and make with the modeling.  So I did.

Color me deeply impressed.  While you’re at it, add some heavy tints of addicted.  I started by modeling our kitchen, and now I want to do the whole frickin’ house.  I’m starting to eye local landmarks for recreation and contribution to the Warehouse and Google Earth.

I don’t have time for this.  I need help.  Stop me before I model again!

Diagnostic Styling

On stage at An Event Apart Chicago, I made reference to recent efforts I’ve been making to develop a set of “diagnostic” styles.  I’d hoped to have them ready for presentation in Chicago, but didn’t get it done in time.

Well, they’re still not really done, but as I’ve now torn them apart and rebuilt them three or four times, with no real end to that cycle in sight, it’s time for me to get them off of my hard drive and into the public eye.  It’s a little bit complicated, so rather than post the whole thing in this entry, I’m going to link to a demonstration page.  I first want to say a few things about it, though.

The primary goal here is to have a set of rules that can be applied during the development phase of a new layout.  These rules’ aim is to visually highlight problems in the markup.  For example, here’s one of the rules:

a[href="#"] {background: lime;}
a[href=""] {background: fuchsia;}

That brings some eye-watering attention to any link that has an empty href, or is (most likely) being used as a JavaScript trigger with no fallback.

Where things got tricky was when I wanted to do things like higlight images that didn’t have alt or title attributes.  In a perfect CSS3 world, I could just say img:not(img[alt]) to select non-alted images.  At least, I think that’s what I would say—:not() syntax makes my temples throb.  Since I was developing these with the idea of releasing them to a more general development audience, though, I decided to use regular old non-:not selectors.

So what I ended up doing, in slightly simplified form, was this sort of thing:

img {outline: 5px solid red;}
img[alt][title] {outline-width: 0;}
img[alt] {outline-color: fuchsia;}
img[alt], img[title] {outline-style: double;}
img[alt=""][title], img[alt][title=""] {outline-width: 3px;}
img[alt=""][title=""] {outline-style: dotted;}

The logic works out like so:

  1. Set all images to have a big red outline.
  2. If an image has both alt and title attributes, set the outline width to zero.
  3. If an image has an alt attribute, set the outline color to fuchsia.  This means the outline of any image that doesn’t have an alt attribute will stay red.
  4. As long as an image has either an alt or a title attribute, make its outline style double.
  5. If an image has an empty alt with any kind of title, or vice versa, make the outline’s width three pixels.
  6. If an image’s alt and title attributes are both empty, then make the outline dotted.

Whew.  Maybe I piled a few too many conditions in there, but I wanted to get some finer granularity on the results, which you can see demonstrated (along with several other things, like highlighting table headers without scope attributes and tables without summary attributes) at the diagnostic demonstration page.  On that page, you’ll see a number of examples and the style sheet that drives them all.  If it’s getting in the way of seeing what’s going on, move the mouse over it to mostly clear it away without actually removing it.  Mouse back out to bring it back.  (Maybe I should reverse those states—what do you think?)

Admittedly, much of what these styles do can be replicated with tools like the Web Developer Toolbar.  The advantage I find with writing my own diagnostic styles is that I can tune them to present exactly the way I want them.  Outlining deprecated elements is fine, but what if I’d rather make them lime-green on cyan to really drive the point into my optic nerves?

Anyway, I don’t for an instant think that these constitute a replacement for the WDT.  They’re just another handy tool to have in the toolbox.

A final note: when actually using diagnostic styles, all of the declarations should be marked !important so as to ensure their application.  I left those directives out of the demo page for clarity’s sake.  If you’re going to use diagnostic styles of this sort in your own projects, remember that you’ll need to add them.

I’m putting these out for comment, suggestions, and general community improvement.  Anyone see things we could add or upgrade?  Let us know!

April 2014