Seeding text files with HTML tags is no fun. We looked at four programs that claim to do the work for you.
Rex Baldazo and Steven J. Vaughan-Nichols
Once, text was static, and it sat in rigid arrays between white-spaced prison bars. Then hypertext and hypermedia unlocked the door, and the World Wide Web pushed it wide open. Now tens of thousands of would-be on-line publishers are rushing through it to create Web documents based on Hypertext Markup Language (HTML). Hot on their heels, software developers are recognizing a golden opportunity to sell tools that make on-line publishing as easy as possible.
Interleaf's Cyberleaf, for example, is a sophisticated file translator that speeds up HTML document production and management. Other programs are little more than collections of HTML macros
stuck on the back of a text editor. In between are tools that vary in complexity -- some of which require expert knowledge of HTML, while others attempt (and fail) to be complete WYSIWYG editing environments.
For this review, we passed over enterprise-level programs, like FrameMaker, which have an HTML component but are designed primarily to handle networked document creation, CD-ROM publishing, on-line technical documentation, and so on. We also skipped minimal HTML-conversion programs (e.g., freeware products for Microsoft Word and Novell's WordPerfect). Instead, we focused on software that truly automates the process of tagging documents for publication in HTML-compliant form.
Two of the four programs we reviewed, SoftQuad's HotMetal Pro and InContext Systems' Spider, provide both authoring (i.e., you can use them to create an original from scratch) and editing (i.e., conversion) of existing documents. Cyberleaf handles only the latter process, but it adds a major feature that none of the oth
er products has: the ability to coordinate and link your HTML documents to create a repository. Brooklyn North Software Works'
HTML Assistant Pro
is mostly for authoring; given its limited import features, it is best used for creating Web pages from scratch.
Most Web editors, including HTML Assistant Pro and HotMetal Pro, are available in freeware versions that contain most of the features found in the professional releases, so you can gain a quick appreciation of each one's strengths and weaknesses. (See
the Product Information box
for the addresses of the vendor Web pages where you can download the freeware.)
One thing's for sure: All these editors are worth the price of admission. While you can
write excellent Web pages
with only a text editor and some expert knowledge of HTML (the HTML encoding is all done in ASCII text), even the lowliest of these products makes it easier to unleash your text on the Web.
The Prob
lem Defined
HTML tags are simply ASCII codes, embedded and visible in the text, that Web browsers, such as Mosaic and Netscape, can interpret as formatting commands. At a minimum, HTML authoring/editing tools automate the tagging process, but this says nothing about the HTML know-how you need to put the tags in the right place.
The better tools provide guidance in this area, insulating you from HTML by querying for decisions such as whether a title is meant to be a section heading, where to place links to other documents, and so on. These tools check a tagged document to see if it's in compliance with HTML rules, and they can even prevent you from making an invalid choice.
We ran the same text file through all four programs and watched them perform the major steps shown in the illustration "From Your Word Processor to the Web" above. In particular, we looked for compliance with HTML versions and extensions, enforcement of HTML rules, interoperability with popular word processors a
nd graphics programs, and usability features, such as an integrated browser and point-and-click commands. The table
"Features of HTML Editing Tools"
is, in part, a checklist of these criteria.
Cyberleaf: Big-Time Web Publishing
Interleaf's
Cyberleaf 1.0
is expensive ($1595) and requires heavy-duty hardware: a Digital Equipment AXP, Hewlett-Packard 700/8xx, IBM RS/6000, or Sun SparcStation 2/IPX with a minimum of 24 MB of RAM and 129 MB of hard disk space. The Microsoft Windows version, due out in the first quarter of 1996, will most likely have a similarly ravenous resource appetite. But the commitment should be worth it -- Cyberleaf is a big-time publishing environment from a company whose Interleaf software is a leader in document management.
First things first: Cyberleaf does not help with the authoring portion of HTML (if you don't count the Home Page Ed-itor, which lets you write new text but provides a minimal set of word pro
cessing tools). Instead, it comes ready to convert Microsoft Word, WordPerfect, Rich Text Format (RTF), Interleaf, FrameMaker, and ASCII files into HTML 2.0 or 3.0 formats. Cyberleaf can also automatically convert a wide array of graphics-file types into GIF format.
With Cyberleaf, you're not just getting a conversion program; you're getting a complete Web development and maintenance environment. When your Web page needs are measured in megabytes and hundreds of links, Cyberleaf is the program to choose.
In our test runs, once it was given the proper marching orders, Cyberleaf had no trouble converting a Word document into HTML. With its workflow-oriented interface, Cyberleaf excels at bringing together documents from disparate word processors and desktop publishing (DTP) programs and quickly and easily converting them into a complete Web document.
Cyberleaf analyzes a document and comes up with appropriate HTML equivalents. For example, by its enabling you to custom-set a Word paragraph sty
le into what you select as the best-matching HTML style, you gain more control over the conversion process. Once you have Cyberleaf set up to your satisfaction, it tears through your files, converting them at a speedy clip.
What puts Cyberleaf into a class by itself is its ability to manage updates to the documents that make up your Web page. Removing a document to which others have pointers can result in broken links to external documents throughout the Web. Cyberleaf identifies such broken links, although it does not have the ability to fix them automatically.
Cyberleaf does, however, automatically maintain intradocument and interdocument hyperlinks in your internal repository, even as you update your documents. Outlines and reusable conversion parameters also help you quickly add new text.
When it comes to graphics, Cyberleaf also kicks rump and takes names. It has excellent graphics-file translators (see
the features table
for a partial list). Our favorite one automa
tically turns any image into a thumbnail-size graphic along with the HTML programming to let viewers opt to see the full-size image.
Not for Pros Only: HTML Assistant Pro
Brooklyn North's HTML Assistant Pro ($99.95) is more of an authoring program than an editor, because it cannot work directly with imported word processor files. The only way that we could use it to get our Word document into HTML was to do it the hard way by saving it in ASCII, thus losing any Word-formatting niceties. In addition, the program comes without a browser of its own. It does, however, make it child's play to call up the browser of your choice to get in-process views of your work.
Pro makes extensive use of toolbars. These are very helpful, if you know HTML. Otherwise, you'll spend a lot of time with the help files. Fortunately, these files are well organized and well written. If you want to learn how to write HTML, you'd be hard pressed to find a better teacher.
Pro has its limitations, ho
wever. It directly supports only HTML 2.0 or lower. If you want to venture into the not-completely charted waters of HTML 3.0 or the Netscape enhancements, you must add these options to the editor. On the plus side, Pro makes it easy to do just that with its option to add HTML tags to the preexisting Users Tool toolbar.
If you need to create pages in a hurry, Pro includes an automatic page generator. It's nothing fancy, but it enables you to make a basic Web page within minutes. After that, you can use Pro's editing tools to fancy-up your raw pages.
Pro's editing utilities, once you know what they do, are easy to use. One nice feature lets you extract uniform resource locator (URL) addresses from Netscape, Cello Bookmarks, and the National Center for Supercomputing Applications (NCSA) Mosaic's INI files to quickly place your favorite Web sites on a page.
HTML Assistant Pro is a good program, but unfortunately the times may have passed it by. Old hands at HTML will find it a comfortable and p
owerful program to use. However, the rapidly growing new breed of HTML users, who want WYSIWYG and drag-and-drop links, will not find Pro to their liking.
Setting Type in HotMetal Pro 2.0
We had hoped to test HotMetal Pro 2.0's ability to import documents from such popular Windows word processors as Ami Pro (now called Word Pro) and Word. Unfortunately, a bug in Windows 95 caused it to balk at just about every document we threw at it. Still, for generating HTML from scratch, HotMetal ($195) is quite competent. It displays HTML tags as icons around the main text (
see the screen
). You can optionally suppress the tags, but, as with older versions of WordPerfect, it's a lot easier to edit with the tags visible.
HotMetal is smart about tag pairs: If you start to select the beginning or ending tag of any pair, it automatically highlights the entire tag. This makes it easy to cut, copy, and paste elements. However, editing tag attributes requires using the no
nintuitive F6 key, not double-clicking or right-clicking to bring up an attribute dialog box. In addition, the editor displays HTML tags inconsistently. When generating a table, for instance, instead of displaying the tag pairs, HotMetal puts up an actual table, and you add your text within the cells of the table.
Unfortunately, the table editor is primitive. No matter how many columns or how much text is in them, the cells remain at a fixed width. Long text lines wrap, making the cells taller; a browser would attempt to make the column as wide as possible to accommodate the text. And there's no way to add rows or columns to a table besides using a text editor to add the lines and then reimporting the new text file into HotMetal.
HotMetal supports some Netscape and HTML 3.0 extensions, but there's no strict way of enforcing which version of HTML you can create. There is, however, an option to check your code, which generates a report detailing the HTML 3.0 and Netscape extensions used in your docu
ment.
Generating an HTML page with HotMetal Pro 2.0 was, on occasion, a frustrating experience. It initially looked like a rich tool set but fell short in implementation. Nonetheless, of the tools we tested, this is the one we're most likely to use to create Web pages. Then we'll move to a text editor to refine and maintain the HTML code.
Spanning the Web with Spider
Spider ($99), from InContext Systems, is really two programs. There's Spider, which is the HTML editor, and Spider Mosaic, which is a customized Mosaic browser. InContext has done some work to integrate the two, but they too often expose their heritage as separate products.
Spider uses a two-pane window to display the page under development (
see the screen
). You can adjust the relative sizes of the panes by sliding the border between them. The left-hand pane is the logical view, showing the HTML tag structure of the document. Icons supplement the tag names, but we found them to be an
noyingly cute after extended use.
The right-hand pane holds the text that appears between the tag pairs. Anchors in the text are bracketed by left and right arrowheads. Editing them is straightforward: A right mouse-click causes a dialog box to pop up with the appropriate attributes. Unfortunately, as with most of these editors, you must understand HTML to know how to apply the correct attributes.
One nice feature is that the Web pages (and their associated links) that you browse through Spider Mosaic can be imported directly into documents that you're editing in Spider. So, if you happen to be browsing, say,
http://www.byte.com/
, you can import any of the links on that page into your own document using the Web Manager option in Spider.
Unfortunately, that's the extent of integration between the two tools. Where Spider Mosaic supports drag and drop to open HTML files, Spider does not. And Spider Mosaic exhibits a problem when it
minimizes under Windows 95: The title bar remains on the desktop. It's more an annoyance than anything else, and InContext is working to resolve the bug.
We also found that Spider balked at reading documents produced by some other HTML editors, because it's finicky about the Document Type Definitions (DTDs) specified in the optional DOCTYPE tag. Deleting this tag is the easiest workaround, or you can edit it to use one of the DTDs supplied with Spider.
In addition, the interface in Spider needs a workaround. There are no tool tips for the toolbars, and given the large number of cryptic buttons on the toolbars, it's difficult to figure out what to do next. The interface has additional quirks, such as not allowing you to delete by using the Delete key (you must type Ctrl-D instead). Ultimately, we have to question the usability of Spider's interface when working on lengthy or numerous Web pages.
The Denouement
The best HTML-conversion tool of the four, irrespective of
price, is Cyberleaf, hands down. While two of the other three products, HTML Assistant Pro and HotMetal Pro, don't even come with a browser for viewing pages, Cyberleaf makes it easy to monitor your progress with full-color, integrated viewers.
Cyberleaf's lack of authoring tools is a design decision, not a shortcoming: Interleaf figures that serious Web publishers will want to use some of the formatting and layout of their existing word processor files, so Cyberleaf comes with the best import features of the group. In addition, Cyberleaf's repository management provides a whole layer of functions that the others don't attempt but are necessary for large-volume enterprises. And it's more automated than the other packages, converting documents with little intervention on your part.
If you're experimenting with Web publishing, have a small number of documents, or plan to use a database manager or other program to oversee your document repository, one of the low-priced HTML editors may be preferable
for you.
Of these, our favorite was HotMetal Pro. Unlike HTML Assistant Pro and Spider, it can import Word, WordPerfect, and Word Pro files, so you can retain more of the value of existing layouts. (The other two don't compensate for this shortcoming by packing word processor features into their authoring environments.) HotMetal lags behind Spider in graphics viewing, but overall, we're more comfortable working in HotMetal.
HTML-conversion tools will evolve rapidly as standards solidify and software vendors scramble to provide the right combination of file compatibility, HTML automation, and document management. These tools are also sure to grow more popular as the demand for Web publishing continues to explode.
PRODUCT INFORMATION
Cyberleaf 1.0.................$1595
Interleaf, Inc.
Waltham, MA
(617) 290-0710
http://www.ileaf.com
HotMetal Pro 2.0..............$195
SoftQuad, Inc.
Toronto, Ontario, Canada
(800) 387-2777
(416) 239-4801
http://www.sq.com
HTML Assistant Pro............$99.95
Brooklyn North Software Works
Bedford, Nova Scotia, Canada
(800) 349-1422
(902) 493-6080
http://fox.nstn.ca/harawitz/index.html
Spider........................$99
InContext Systems
Bethesda, MD
(800) 263-0127
(301) 571-9464
http://www.incontext.ca
Cyberleaf HotMetal HTML Asst. Spider
Pro Pro
=====================
=========================================================
HTML Features
HTML versions supported 2.0, 3.0 0.9, 1.0, 0.9, 1.0, 0.9, 1.0,
2.0, 3.0 2.0, 3.0 2.0, 3.0
Supports Netscape HTML N Y Y Y
extensions
Enforces HTML rules Y Y Y Y
Supports forms creation Y Y Y Y
Templates included Y Y Y Y
WYSIWYG view of entire Y N(1) N(1) Y
Web page
Preview graphics in Y N(1) Y Y
WYSIWYG form
Integrated Web page viewer Y N(1) N N
Number of file formats 30 30+ 4 1 (HTML only)
supported
Major file formats ASCII, HTML, ASCII, HTML, ASCII, HTML, ASCII, HTML,
GI
F, JPEG, GIF, JPEG, GIF, JPEG GIF, JPEG
RTF, Post- RTF, AmiPro,
Script, Frame- Word,
Maker, Word- WordPerfect
Perfect,
Interleaf
Can set HTML preferences Y Y N Y(2)
during import
Automatic insertion of Y N N Y
URL addresses
Word processing features
Word processor formats Word, Word- Word Pro, None(3) None(3)
supported perfect, Inter- Word-
leaf, Frame- Perfect
Maker
Spelling checker N Y Y Y
Thesaurus N Y N N
Supports macros Y Y Y N
Standards-compliant table Y Y Y
Y
editing
Provides document Y N N N
management
General Features
OSes supported Unix, (SunOS, Windows Windows 3.x Windows 3.x
Solaris, HP/ 3.x, Mac-
UX, AIX, OSF) intosh
Minimum RAM requirements 24 8 4 4
(MB)
Hard disk requirements 65 15 1 7
(MB)
KEY:
(1):
Available using a separate browser.
(2):
Not a menu choice, but available via Spider's
DOCTYPE DTD utility.
(3):
Can import ASCII files exported from word processor.
Y:
Yes
N:
No
screen_link (146 Kbytes)
HTML-conversion programs either import plain text files, in ASCII or a popular word processor format
(above)
, or provide a simple word processor for writing text from scratch.
Standard browsers read the HMTL coding and display the finished Web page
(right)
.
You then employ menus and pick lists to choose high-level functions (e.g., make header or align paragraph) or to directly insert the equivalent HTML tag
(left)
. The underlying HTML coding is pure ASCII.
screen_link (30 Kbytes)
Cyberleaf has the most graphical interface of the programs we reviewed. It offers easy point-and-click access to files in a document repository.
screen_link (37 Kbytes)
Like HotMetal Pro and Spider, HTML Assistant Pro offers drop-down menus for choosing HTML tags.
screen_link (71 Kbytes)
HotMetal Pro
(left)
lets you choose clearly worded formatting options from a pick list or from the toolbar. It then inserts iconized HTML tags while reformatting the text to be closer to its final form.
In contrast, Spider's automated tags
(above)
appear in a separate window and can get out of alignment with the text, although the right-hand window is more representative of the Web page.
Rex Baldazo is a BYTE technical editor who works on the magazine's Internet publishing ventures. Steven J. Vaughan-Nichols is a freelance writer and consultant who specializes in Internet and other communications issues. You can contact them on the Internet at
rbaldazo@bix.com
and
sj
vn@ichange.com
, respectively.
Web Publishing Made Easier
