A recent post on the TechNet mailing list posed an interesting question [ed: TechNet is now Claris Community. The link has been updated accordingly.]:
When displaying an image in a web viewer, is there any way to scale the image to fit?
Of course container fields have this ability built right in (by way of the Format → Graphic command). But there are several reasons you might want to show pictures in a web viewer instead. Most notably, the images might already be on a web site. There’s no reason to copy them into the FileMaker database too, wasting space and adding more to your already busy schedule. Is there any way to give web viewers container-like scaling ability?
This question inspired me, so I put together a solution that can be used generically. The result is a simple web page that can, using Javascript, load an image, figure out how big it is, and scale it appropriately. First, here’s the source of the page. If you’re curious about how it works, read on.
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html>
<head>
<style>
body {
border: 0;
margin: 0;
}
img {
display: block;
margin: 0 auto;
}
</style>
<script type="text/javascript">
function resize_image() {
var img = document.getElementsByTagName('img').item(0);
if (img.offsetHeight > img.offsetWidth)
img.style.height = '100%';
else
img.style.width = '100%';
}
</script>
</head>
<body onload='resize_image();'>
<script type="text/javascript">
document.write(
"<img src='" +
location.search.replace(/^\?/, "") +
"'>"
);
</script>
</body>
</html>
To use this in your own systems, copy the source code into a new file. Name it “scale-image.html” (or anything you want) and put it on your web server. To show a properly scaled image in a web viewer, you load this web page, and tell it what image to load in the query string (in other words, add a question mark and then the image URL to the end of the page URL). Sounds complicated, but really it isn’t. Here are a few examples:
# Load an image with a complete URL:
http://me.com/scale-image.html?http://foo.com/path/to/image.jpg
# If the image is on the same web server:
http://me.com/scale-image.html?/path/to/image.jpg
# If the images are in the same folder as the page:
http://me.com/scale-image.html?image.jpg
Of course you can use any image type your browser supports: You’re not limited to JPEGs. Finally, note that this page only works properly if the web viewer itself is perfectly square (ie: 120px by 120px or 300px by 300px). Use the Object Info pallette (View → Object Info) to easily size your web viewer. If there is enough interest, I can revise the code to work in web viewers with any aspect ratio.
How Does it Work?
If you’re new to dynamic HTML the web page may seem a little mysterious, but it is actually pretty easy to understand if you are familiar with HTML and Javascript. Let’s dissect it.
[ed: This may be dynamic and HTML but it is not Dynamic HTML in the formal sense. I think I just didn’t know the details well enough when this was posted.]
First, you see a pretty typical top-of-the-page. We declare the version of HTML we’re using, and start the page as normal.
The first important bit is this:
<style>
body {
border: 0;
margin: 0;
}
img {
display: block;
margin: 0 auto;
}
</style>
This embedded CSS stylesheet tells the web viewer a few important things:
The
body
part ensures the page body has no margin or padding. In other words, we want the contents of the page to push right up to the edge of the web viewer, with no gap. Normally, page content is indented a little on all sides, but we want our image to fill things up completely.The
display: block
part tells the Web Viewer to treat images as block elements. This makes it possible for us to size and position the image precisely.The
margin: 0 auto
part means the image should start at the very top of the web viewer (no top/bottom margin) and it should center itself right-to-left (automatic left and right margin).
Next, we define a Javascript function. This code doesn’t run right away, but later it will be called. It does the actual work of resizing the image. Here’s the function:
<script type="text/javascript">
function resize_image() {
var img = document.getElementsByTagName('img').item(0);
if (img.offsetHeight > img.offsetWidth)
img.style.height = '100%';
else
img.style.width = '100%';
}
</script>
In CSS, if we set just the width of the image to 100%
then the image will automatically be scaled so it is exactly the width of the web viewer. If you don’t set the height at all then the web viewer will keep the image’s aspect ratio. In other words, it will let the image be as tall as it needs to be so it doesn’t look squished. Likewise, if we set just the height to 100%
, the width will scale accordingly. All we need to do is decide if the image should be just as wide as the web viewer, or just as tall.
[ed: CSS has come a long way since this was posted. All this information is very out of date. Today you can simply use max-height: 100%
and max-width: 100%
on the image and the browser will do the right thing. No Javascript is needed here at all.]
If you think about it for a minute, you’ll agree that this depends on whether the image is taller than it is wide, or wider than it is tall. A tall image will fill the web viewer from top to bottom, and won’t be quite as wide as the web viewer itself. A wide image, on the other hand, will fill it from right to left, and be a little shorter.
So this function first get ahold of the image, storing a reference to it in a variable called img
. It looks at the offsetHeight
and offsetWidth
of the image (these are just funny ways to refer to its height and width). If the height is bigger, it sets the height to 100%
. If the width is bigger, it sets the width to 100%
instead.
Remember though that this function is only defined at this point. We haven’t actually run the code yet. The reason is simple: the image doesn’t exist yet. Next, the HTML looks like this:
<body onload='resize_image();'>
<script type="text/javascript">
document.write(
"<img src='" +
location.search.replace(/^\?/, "")
+ "'>"
);
</script>
</body>
In this very short HTML body, we use another Javascript snippet. This time we generate an HTML img
tag with a src
attribute that refers to the image we’re supposed to load.
The image URL itself is loaded from the location.search
property, which is just Javascript’s way of fetching the junk after the ?
in the page URL. Unfortunately, the value in location.search
includes the ?
itself, so we use the replace
function, with a regular expression to remove it.
Now we have a web page with an image in it. If you look at the very top of the last code snippet, you’ll see where we call the resize_image
function: in the body
’s onload
handler. This tells the web viewer to run the Javascript resize code only after the page (and the image on it) have finished loading.
When the function runs, it adjusts the size setting for the image, and the image is properly scaled. Luckily, Javascript is very quick, so all you see is a perfectly scaled image from the start.
This sample was thrown together quickly, so it has room for improvement. To wit:
The web viewer has to be square. It would be better if it inspected the page size and adjusted its decision making accordingly, so it could scale an image in any size web viewer.
The image is centered left-to-right, but not top-to-bottom. Centering vertically in CSS is a bit of a nuisance, but there are several ways to do it, like this, this, and this. I think for the sake of simplicity, I would look at two approaches: Calculate the appropriate margin in Javascript, and set it; or just put the whole thing in a single-cell table. [ed: Or wait five of sixe years for CSS to catch up with reality. Vertical centering is now easy in CSS.]