[]

[]

Introduction

When I first started surfing the World Wide Web, I wondered over and over why I didn't see pages that were more like magazine articles. There were some pretty ones, sure, but few had multi-column text or anything like what you'd see in a magazine. When I began designing my own, I started to see why. I was under the mistaken impression that HTML was a page layout language, an impression I feel all too many web authors still hold.

I put up my first home page, a glossary of commonly used terms in Doom level design (a revamped version is still available), using the knowledge I had gleaned from viewing the pages of others. I thought my page looked great, and so did a few other people. Unfortunately, I now realize that I tried to add so much to my page that it really looked anything but great.

Meanwhile, I had read a fascinating article by Larry Aronson in the October '95 issue of NetGuide. He spent a brief while talking about the clash between NS standards and the standards put forth by the W3C. He said that there were many who "questioned the values of standards altogether, declaring that the market should rule." When I first read this, I was one of them.

I started spending some time on comp.infosystems.www.authoring.html, where I read some posts by people such as Mr. Alan J. Flavell, Abigail, and others, who argued against using balkanizing NS extensions. I tried to justify my use of NS extensions until I realized that they had a significant point. The Web was about presenting information -- whenever you start dictating how this information is to be presented, you limit your audience. The statement from the article began to lose its appeal.

Now we come to my current position. I'm for a WWW where anyone, with any computer (whether it be a 366 MHz DEC Alpha or a simple text-based terminal), can log in and receive information that their browser can process. As such, I'm against tags that control layout and leave no alternative for text-browsers (a severe example of this is NS' client-side image mapping). I'm also against misuse of tags simply to force a certain layout (i.e., using <UL> or <DD> out of context to force an indent).

At first all this may seem unrelated to our topic. It isn't.

The meaning of WYSIWYG

What I tried to lay out above was the foundation for what I'm about to say. THERE IS NO SUCH THING AS A WYSIWYG HTML EDITOR. HTML was meant to be, and IS, a language for marking up the structure of pages, not the layout of pages. As such, no editor can guarantee that what you produce will look the same on anyone's computer but yours. This doesn't mean that documents that allow you to display and preview at the same time are useless. Said Dan Connolly, one of the fathers of HTML, in a post to comp.infosystems.www.authoring.html

Granted, WYSIWYG is a misnomer. I prefer the term "direct manipulation," meaning that you edit a natural representation of the information (multiple fonts, colors, etc.) directly -- with the mouse, keyboard (or voice, ...)

I share the sentiment. WYSIWYG editors are also particularly useful on Intranets, where you know exactly who has access to your pages as well as their system configuration. You can design your site with little regard for how correct your HTML is for the simple reason that you're working in a controlled environment. Here editors like AOLPress, Atrax and Netscape Navigator Gold prove particularly useful. Here I'll talk about each and summarize what's important.

The Editors

[] []

Netscape calls Navigator Gold the perfect tool for Intranets, and with good reason. After all, as I said before, Intranets are particularly easy to design for because you know exactly who'll be viewing your site, what their system configuration will be and what browser they'll be using. Internet design is another story, however.

Netscape Navigator Gold was designed to allow you to edit your page just as you'll see it; it does this job remarkably well, though neither image alignments nor forms have been fully implemented yet. Image alignments are available within the editor, but all images are shown inline -- to see how your page actually looks, you have to load it in the browser.

Less care was even taken with forms. View a page containing them and you'll see the symbols for unrecognized tags where dialog boxes and buttons should have been.

As those who've known me will vouch for, I haven't been particularly too fond of Netscape in the past. It's not that I don't like their browser -- it's probably the best one available for my platform (win16)-- but rather that I don't like their philosophy. Despite their statement of support for open standards, many of their extensions have balkanized the Web. Alas.

However, I actually like Navigator Gold. It's easy to use, fast and allows you to input raw HTML for tags that it doesn't understand yet. It does have a dark side, though: It promotes the use of extensions that haven't yet been standardized and the misuse of tags that have.

[]

From the distorted toolbar above, you'll notice the Indent buttons, which increase and decrease the indent of paragraphs one level. Well, since I wasn't aware of an existing HTML tag to do this, I tried this and looked at my source. It turned out that NS merely put <UL> and </UL> tags around the part of the document in question! Not only would this not validate, but it used a standard tag in a way that older browsers could choke on.

What I find ironic is that Netscape will allow, even encourage, this misuse of HTML. After all, way back when they were implementing some of their first NS extensions, they chose <CENTER> over <P ALIGN=center> partly because " <P ALIGN=center> breaks older browsers when P is used as a container". Surely using UL out of context does far more damage! Also notable is that NS changes <em> to <i> and <strong> to <b> in any document it imports. Once again, this is fine for most Intranets, but those wishing to design for the Internet should take more care to make their documents more accessible.

Other than those (and a few more) violations, NS Gold generates passable HTML. With a few modifications, and a decent amount of time, one can get HTML produced by NS Gold to validate fairly easily.

Atrax: WinWare Inc.'s offering

[]Atrax is a recent editor, put out by WinWare Inc., that works on a drag-and drop principle. You select items from a list and drag them into your project file, and then modify their content and properties. This works fairly well, although it takes longer to learn to use Atrax than either AOLPress or NS Gold.

After you learn to use it, Atrax is fairly capable. It includes a variety of tags (including FONT and some of its attributes) and allows you to insert your own raw HTML, if you wish. It also comes with several built-in templates, although when I beta-tested it, nearly all were under construction. Also notably lacking is the ability to import already completed HTML files.

After you've finished your project file, Atrax will then have you compile it, which is the step that generates the HTML. What I find uncanny about it is how deviant the final HTML generated is from the standard! At first glance, the source looks a little jumbled, with the attributes for the <BODY> tag on a separate line from the tag itself. You then notice that all the elements of your document are enclosed in tables! Examining the table cells, you see that when you dragged and dropped in a header, all it did was up the font size.

Again, this may be fine for Intranets, but when you're trying to make your document available to a wide audience, you want to be sure to markup structure, not layout.

[]

AOLPress is the name of the software formerly known as NaviPress or GNNPress -- it's essentially an easy-to-use HTML editor based on the same principles that guided NS Navigator Gold. I tried NaviPress 1.1 a while back and it became a near-permanent fixture on my already-bloated hard disk.

First off, the bugs: AOLPress is slow. It's not all that sluggish, but when you have many images on the screen, it's easy to notice the delay between when you type something and when it actually shows up on screen. That's about the only problem that easily comes to mind.

What really impressed me about AOLPress is that it was the first WYSIWYG editor that really produced compliant HTML. You can set it for NS HTML, in which it allows certain proprietary code, or change it to HTML 3.2, the current proposed standard, in which it's more strict.

To test it, I loaded up a page I had been working on (one I thought compliant), added a space or so, and sure enough I got a dialog box saying that my page had been modified for "HTML conformance." I thought, " aha! It's changed my compliant HTML into something horrendous." I looked at the source and was pleasantly surprised -- while it may have made some small changes, they probably didn't stop it from validating, and the source looked much more structured. It indented list items and produced much more attractive HTML. About the only problem was that new documents are not started off with a DOCTYPE declaration.

AOLPress also understands more tags than the others. It doesn't treat convert bold to strong text or italic to emphasized text, but treats them separately (it puts strong text in red by default, I believe). I like this; it reinforces the difference between structural markup and presentational markup.

AOLPress, as an editor, aligns images the way they'll appear in the final document, unlike Netscape. It also understands forms and has its own tools for building them! A great tool for Intranet or Internet design.

Conclusion

The real problem with WYSIWYG editors is that many people design their pages around them. This is fine for Intranets, where you know that others will be using the expected browser; for others, however, receiving a page which uses <UL> to force an indent may simply be confusing.

If you're doing Intranet design, any of these packages will work, my preference going to NS Gold or AOLPress; if you're into WWW design, and you'd like to produce compliant pages, my preference goes to AOLPress. It's the most WYSIWYG of any featured and it's simply produces HTML that's almost validator-ready.


[] Written and maintained by Paul Pollack; last updated 1 Oct. 1996

paulp at gte.net