“Bread crumbs”(perhaps how you got here)

Those that have followed this Web site will undoubtedly have noticed that its appearance changes from time to time. In this, it resembles other Web sites.

But, unlike the simple whims of occasional redesign, whims that keep Web designers in business, this Web site has changed because the software used to create the site ceased to work in August. That software was, of course, Trellix (aka “CuteSITE Builder”, complete with the bizarre and entirely unnecessary capital letters). In June there was a complete rebuild of my computer system. It now consumes enough electricity to be noticed in our electrical bills. This has the pleasant effect that in the winter the study, formerly the coldest room in the house, has become the warmest.

In doing the system rebuild, I strongly suspected that I would lose Trellix. The reason for this was that the few and mostly minor changes made by GlobalScape, the company that renamed Trellix so strangely, included above all changes to the proprietary format in which the source files were stored.

At this point I should comment briefly about the way in which Web pages get created, sharply distinguishing the pages from the “Web sites”.

I first began to teach about HTML (and such other mark-up languages as existed at the time) in 1992. At that time, HTML (the mark-up language in which Web pages were expressed) was very simple, and most people that wanted a Web page wrote the code by hand, perhaps using Microsoft’s Notepad or Unix’s vi. The crucial feature that was needed was that the editor had to produce a pure “character” file. (And at that time, “character” meant the 96 “printable” ASCII characters).

Also in those early days, Web sites were generally small, confined to a few pages. One could teach all there was to know about HTML and Web sites in general in about two hours. I had quite a few students that made a bit of pocket money by using their knowledge to produce Web sites for the adventurous businesses that wanted to get in early on what was clearly going to be a major change in marketing.

In those days, developers typically wanted the entire Web site, code and all, to be conveniently located on a disk (or, more likely, a single floppy diskette).

But progress occurred. Web sites grew bigger and bigger and sprouted more and more features. We soon saw Web pages in bright colors, with animated graphics, and interactive forms. It became difficult for the Web designer that used “hand-coded” HTML to keep the appearance of the Web site consistent. There also arose the ancient software-development problem of version management.

I do not know when or how the general solution to these problems developed. That solution was to bundle together all of the site’s components into a single (compressed or archived) file; about the time this solution became popular, “WYSIWYG” editors also appeared on the scene - sometimes in garb that made them look very much like traditional word processing software.

This pair of developments, going along hand in hand with each other, profoundly changed Web development. In principle, “anyone” could now develop a Web site. The new software packages of the late 1990’s (among them Trellix, Net Objects Fusion, Macromedia’s Dreamweaver, and Microsoft’s Front Page) provided the WYSIWYG editor and the site management software that “simplified” development. (Whether there was real simplification, I’m not so sure.)

There were two problems. The first (and much less important) was that the editor had to write HTML, since the protocols used to transmit Web pages and their contents were still fixed on HTML (although HTML itself was always in a continuous state of evolution). The need for the editor’s services in writing the HTML that constituted an actual Web page resulted in an unfortunate problem - the quality of the HTML produced was often very poor. Most of the time this was not a huge problem; but often, as in the case of subtle or dramatic effects, it was necessary for an “expert” (i.e., someone who actually knew HTML and could edit the source code with good old Notepad or another similar editor) to modify the code. When the HTML was badly written (as was the case with many early packaged Web-development systems) the cost of Web development work rose significantly.

The second problem was that there were no standards for the portmanteau files that stored the entire Web site. Every software house that was in the business had its own proprietary format.

And, in the case of Trellix / CuteSITE Builder, the proprietary format kept changing. Brilliant design work was done on some of the Web development software. Elegant interfaces began to appear - even now, I admire Front Page and Net Objects Fusion. But the brilliance of the user interface concealed the dirty secret, that there were no standards. A Web developed using Dreamweaver could not, for example, be ported easily and quickly to Net Objects Fusion. This remains the state of affairs, even today.

As a result, when I finally gave up on Trellix / CuteSITE Builder, I had to reconstruct the entire Web site, from the ground up, in my new development platform. This is, of course, insane.

In word processing, by contrast, almost every known word processing system can read and write files in any of several common formats. Spreadsheets are similar. Database systems usually provide easily-used transitioning tools. But I know of no software that allows one to take a Trellix Web site and port it to Front Page. In fact, there seems to be no software that even allows convenient and natural transitioning from one version of Trellix to the next. This state of affairs is simply unacceptable.

Of course, I could go back to hand-coding everything. It’s not really that hard to do, although HTML (and its daughter, XHTML) is now a lot more complicated than it used to be. But when one tries to hand-code a large Web site, the result is always unsatisfactory (coding error rates rise) and the time spent developing pages becomes a serious problem. The consequence is that today most Web sites are developed in some proprietary format.

If I were God and able to rearrange the affairs of the World conveniently and wisely, I would make all proprietary file formats illegal. Alas, I’m not God, and proprietary file formats continue to proliferate.

And, of course, the matter of porting a Web from one development environment to another is a serious problem. But there is an advantage: by confining the developer to a single format, incompatible with others, the software publisher tends to keep the customers already committed to that format. The cost of transitioning is too great for most developers and their customers. This in turn means that Web sites are developed in a manner that is, when viewed globally, inefficient and ineffective. We have beautiful Web sites with fancy features, and they are absurdly costly. Welcome to the Timid New World.

---

Note added Sun, Dec 13, 2009:

Well, now I’ve built three Web sites, including this one, using NetObjects Fusion 11. I am profoundly unimpressed. Compared to Trellix, it takes roughly twice as long to edit a page, because the affordances are stupid poorly designed and drag-and-drop does not work correctly all the time. Compared to Trellix, NOF does a poor job of maintaining the uniformity of page layout, and it entirely lacks Trellix’s handy instant navigation buttons.

The site navigation window that opens as one of the tool bars lacks all the features of the “Site” view; the Site view itself by default employs an idiotic “organization chart” top-down style that is useless on a site with more than about 20 pages. The alternative “outline” version of the Site view is better, but poorly integrated with the page-detail section that is supposed to supplement the view of the items (pages) in the outline.

Trellix was always distressingly slow when editing a large page, and it bogged down on large sites (as large as this, anyhow; in the great run of things, this is not a huge site). NOF is worse in both respects. Its editor is astonishingly deficient in even the simplest features, such as regular expressions in search-and-replace. It takes an astonishing amount of time to post this Web site (roughly 2 hours at present on a medium-quality broadband connection), and the program is highly unstable, especially when posting. On a computer with 6 GB of RAM and an Intel chip with 8 CPUs, NOF is very slow.

The HTML code produced by NOF is not very good, but it’s not bad. On the other hand, integration of an outside HTML editor is almost non-existent (it’s supposed to be there, but it’s not).  Integrating forms is also difficult, although the form designer is pretty good.

On the other hand, NOF produces nice-looking pages, and its “assets” are fairly easy to modify, better than Trellix was.

All in all, while I’d give Trellix a grade of B as a Web-site editor, I’d give NOF a grade of C-. I’m already looking around for something to replace it - something affordable, that is. Maybe I’ll trot out my old copies of Front Page and Dreamweaver and see if there’s anything there. Or maybe Namo .. . but the message to NetObjects should be this: I’m not happy.

 All text on this page is the work of J W Durham and is licensed only under terms of the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 United States License. Other licensing terms may be available. E-mail me