Previous Issue | Search TidBITS | TidBITS Home Page | Next Issue
Is text dead? Not at all! In this issue, we look at Palimpsest, a tool for managing large volumes of text; CopyPaste, a multiple clipboard utility; and Natural Order, an extension that (finally!) sorts text and file names like a person would. Also in this issue, readers respond to the new crop of HTML editors, Speed Doubler gets an important update for Mac OS 7.6, Be stops making the BeBox, and Apple announces major internal changes, price cuts, and a Rhapsody kernel.
Copyright 1997 TidBITS Electronic Publishing. All rights reserved.
Information: <firstname.lastname@example.org> Comments: <email@example.com>
This issue of TidBITS sponsored in part by:
Aladdin Systems -- 408/761-6200 -- <http://www.aladdinsys.com/>
Makers of StuffIt Deluxe 4.0, the Mac compression standard, and
InstallerMaker 3.1.1, the leading installer for Mac developers.
Apple Price Cuts -- Last week, Apple announced it had lowered dealer prices as much as 27 percent on a wide range of Macintoshes. The largest cuts appear in the mid-to-high end of Apple's desktop Power Macintosh line, with prices reduced as much as $1,000 on Power Mac 8500 and 9500 models; also, the high end of the Performa 6400-series has been discounted 15 to 18 percent, and Apple's Workgroup Server 7250 and 8550 series are discounted 7 to 13 percent. Apple is no doubt attempting to boost sales volume after an unspectacular holiday season and clear inventory in anticipation of new models to be announced later this month. [GD]
Apple Layoffs and Reorganization -- It's been just over a year since Apple's last reorganization and major round of layoffs, but Apple is expected to announce another sweeping set of changes intended to further focus its business model and cut operating costs by 25 percent. According to reports, Apple plans to organize around three major markets (publishing, education, and the Internet), sell off some key assets (including Pippin and possibly the Newton division), and lay off 2,000 to 3,000 employees this year. In January, Apple said it would announce restructuring plans aimed at cutting some $400 million in costs. After losing about $900 million during the last year, analysts generally agree Apple must reduce its operating costs to hope to return to profitability. Bandai says it will continue to upgrade the Pippin platform and introduce it to new retail markets in 1997. [GD]
Mach Speed -- In a letter to developers last week, Apple's Chief Technology Officer Ellen Hancock announced that Apple has decided to use the Mach kernel as the foundation for Rhapsody, although no other details were given, including which version of Mach Apple plans to use. OpenStep is targeted at the Mach 3 kernel, whereas NeXTstep uses a variant of Mach 2.5. Presumably, Apple chose Mach in order to bring Rhapsody to market more quickly.
The Mach kernel was originally developed at Carnegie Mellon University, and is currently used by many environments, including IBM's AIX, Apple's MkLinux project, and Tenon Systems' MachTen Unix environment for the Mac. Though most operating systems using Mach have been based on Unix, that doesn't necessarily mean Rhapsody will have Unix at its heart. [GD]
Speed Bump for Speed Doubler Users -- Connectix has identified potentially serious problems with using Speed Doubler and Apple's Find File under the new Mac OS 7.6. Problems range from innocuous ones like Find File not displaying information to scary ones like disk catalog errors. Last Friday, Connectix released the Speed Doubler 2.0.1 Updater, which corrects the problem for 2.0 users. Speed Doubler 1.x users should be able to update to version 1.3.2 in the very near future; Connectix may have posted the 1.3.2 Updater by the time you read this text. These releases are both U.S. versions; Connectix plans to release localized versions soon. [TJE]
Connectix -- 800/839-3632 -- 415/571-5100 --
415/571-5195 (fax) -- <firstname.lastname@example.org>
No More Be Hardware -- Industry darling Be, Inc. announced last week it will stop making its own hardware line - the BeBox - and focus purely on developing the BeOS for PowerPC-based Macintosh computers. Be points out (rightly) that it's difficult for a 50-person company to design hardware and an operating system, and the lion's share of their target user and developer markets are already using Power Mac hardware. Be promises to support current BeBoxes for at least the next three years. [GD]
Updated QuickMail Pro -- CE Software recently announced the release of QuickMail Pro 1.0.1, which fixes several bugs and improves a few features in the company's POP3 email client. An updater for Macintosh users is available at the company's Web site, and CE expects to release an updater for Windows later in February. [MHA]
Rev Now Has Online Ordering -- The folks at 6prime wrote to say that they were inundated with orders for Rev after my review of their excellent revision control program in TidBITS-362. To better handle the volume, they've scrambled to put an order form online at the URL below. The price remains $49.95. Shipping and handling (and a bonus mug) is $4 for U.S. and Canadian addresses, but if you don't need a disk or mug or are outside of the U.S. or Canada, you can save the $4 and go for email delivery. [ACE]
by Tonya Engst <email@example.com>
In TidBITS-362, I wrote about how several upcoming HTML editors use tables or Java to offer free placement of objects. Several readers responded with comments about problems with the Web pages those editors are likely to produce, and with thoughts about where this trend may take us.
Bill Seitz <firstname.lastname@example.org> noted:
Lots of pages on the Web look stupid to me because I set my default font to Palatino 12 instead of the tiny and ugly Times 12. Cascading Style Sheets offer some additional placement control without resorting to tables, but still target their features to publishers attempting exact control over the user's view. I sometimes think these people should just make a giant JPEG for each page and stop the pretense.
Brad Kemper <email@example.com> chimed in as well:
I think free placement is a disturbing trend, not because of the code it produces (I would like not to be concerned with code at all), but because fixed-width pages do not account for the specific needs of people who read text onscreen. Since the first Macs, text has automatically wrapped to fit the size of the window. Now, thanks to programs that create Web pages with items placed to exact pixel coordinates, we lose this capability. Perhaps it is because we are using a page paradigm instead of thinking of Web pages as windows or screens of information. We've taken a huge step backwards. We should take a hint from people who design interfaces for computer programs: good design for monitors is different from that for print.
Sajid Martin <firstname.lastname@example.org> worried about speed, commenting:
I think an important disadvantage could be that using tables to configure the entire page results in much longer page rendering times, and slows down scrolling in a browser. But, I think the trend to make coding - including scripting - unnecessary may be good in the long run for the end user.
by Adam C. Engst <email@example.com>
Here's a silly question. Are the digits between one and nine represented by a single character, or by a string of characters? In other words, when you type the number one in a filename, do you always prefix it with a zero? In all likelihood, a number of people are nodding their heads and thinking, "But of course I do that, otherwise files with numbers in the names don't sort right."
We fought with this problem with TidBITS in our early years, because although I was clever enough to prefix my single digits with a zero to pad them into double-digit numbers, I never imagined that TidBITS would be around long enough to hit TidBITS-100, necessitating a mass renaming of the first 99 issues to include an additional leading zero to pad everything into triple-digit numbers. And even today, we can't imagine that we'll be doing TidBITS long enough to need to go to quadruple-digit numbers (although stranger things have happened).
In case you're still wondering what I'm blathering about, here's an example. Assume you keep your own collection of TidBITS around, but for some reason, you rename them with more natural names. You might have the following problem with the first 12 issues when viewing them in the Finder.
TidBITS-1.html TidBITS-10.html TidBITS-11.html TidBITS-12.html TidBITS-2.html TidBITS-3.html TidBITS-4.html TidBITS-5.html TidBITS-6.html TidBITS-7.html TidBITS-8.html TidBITS-9.html
The trouble should be obvious - why do issues 10, 11, and 12 sort before issue 2? It's because the Macintosh System is messed up, that's why. Basically, way back when, someone at Apple decided to implement sorting the same way that Unix, DOS, and everything else does, by comparing one character at a time, instead of taking into account the fact that numbers don't sort in alphabetical order the way words do.
For years, the universal solution has been to prefix single-digit numbers with a zero, so they sort before double-digit numbers. As the number of digits increase, so must the number of padding zeroes. Computers should fit the way humans work and think. Instead, when it comes to sorting, humans must fit the way computers work.
No longer. Stuart Cheshire <firstname.lastname@example.org>, author of the popular network tank game Bolo, has released a tiny freeware extension called Natural Order that overrides how the System sorts numeric parts of strings. I won't repeat the example list above, but once you install Natural Order (drop it in your Extensions folder and reboot), issues 10, 11, and 12 would sort properly to the bottom of the list.
Programs that benefit from Natural Order immediately include the Finder (for "View by Name" windows), Standard File Dialogs (in any application), and the Chooser (for sorting lists of zones, servers, and so on). However, Natural Order works by overriding the System's built-in string comparison routines and only benefits programs that call those routines. A number of programs implement their own sorting mechanisms, so those programs don't benefit from Natural Order. A few recent programs (including Anarchie 2.0 and Fetch 3.0.2 and later) include Natural Order's sorting routines internally so they sort sensibly even without Natural Order installed.
Stuart wants to hear about any programs that do not benefit from Natural Order's sorting routines, so he can try to convince the programmers of the benefit of making their programs call the System's built-in string comparison routines (and thus take advantage of Natural Order).
So, if you're bothered by the way the Mac sorts numeric strings, give Natural Order a try. You probably won't even notice it until you see a folder containing numbered files.
by Matt Neuburg <email@example.com>
Today I found myself in one of those situations where I had to carry several separate pieces of information from one application to another. I was building (in Symantec Visual Page) a Web page composed of quotes extracted from Web pages (in Netscape Navigator). For each quote, I needed the title and author (displayed at the top of the page), the extract (somewhere in the middle), and the URL (from Navigator's Location field). Now, how many times do you think I had to switch between applications to create each page? Wrong! For each Web page, I only had to copy the information from Navigator and switch to Visual Page once - carrying with me the three pieces of information in three separate clipboards.
In the middle of this operation I suddenly became conscious of how wonderful it was to be able to do this, and had to stop and dash off this praise of the extension which gives me not just three but ten clipboards - CopyPaste 3.2.2.
Originally, I scoffed at CopyPaste, feeling about it as I once did about drag & drop (crusty, old-timer, Gabby Hayes voice): "Why, for years I've been cutting and pasting one thing at a time, and it's always been good enough for me!" Besides, early versions of CopyPaste crashed certain applications on which I rely. But CopyPaste's compatibility has improved tremendously; and now that I use it, I use it constantly and automatically.
The basic functionality of CopyPaste is simple to describe. You have ten system-wide clipboards, numbered zero through nine. To copy the current selection into clipboard 7, instead of pressing Command-C, you press Command-C-7 - without releasing the Command key until after you've typed the 7. The same interface applies to both cutting and pasting, with Command-X and Command-V. Or, you can use the Edit menu, which CopyPaste provides with hierarchical menus leading to the ten clipboards, even showing a little snippet of what's currently on each one.
CopyPaste provides some nice extras too. You can archive the clipboards as files (one at a time or all at once), and you can have clipboards automatically archived at shutdown and restored at startup. You can copy the current selection to an append file, an option that - for instance - works well for compiling a download list while reading the Info-Mac Digest. And there's a windoid that shows you the full contents of each clipboard and lets you swap clips with one another.
Other functions do not interest me as much. There's an application-switching shortcut I never use because it interferes with HyperCard, and something called Tag and Drop that I don't even understand. Plus, there are a host of clipboard-massaging functions that basically reinvent the Clipboard Magician wheel. Luckily, you can turn off unwanted features in a Preferences dialog, but personally I find this "don't know when to stop" style of programming annoying and ill-advised. What's needed is a component approach so code for undesired functions never even loads.
There are also problems caused by the fact that CopyPaste is an undeniable hack. These troubles change from revision to revision: partly they involve compatibility, but the main difficulty at present is that you lose access to an application's own internal scrap mechanism: every time you copy, even if you just press Command-C, you copy via CopyPaste. This can be problematic for some applications, which may switch to the Finder and back (losing time while windows and palettes are taken down, and more time - and perhaps information - while the contents of the clipboard are converted away from an application's internal format). I would much prefer Command-C just performed the application's original default copy.
Nevertheless, the spirit of CopyPaste is commendable, and - once you've experienced its utility - you can't fathom why it hasn't been built into the System for years. I'd sacrifice all the touted functional improvements of Mac OS 7.6 if only it included ten clipboards. Meanwhile, at a shareware fee of $20, CopyPaste is a bargain-priced way to give your machine a new soul.
by Matt Neuburg <firstname.lastname@example.org>
Readers of TidBITS know of my unabashed obsession with the storage and retrieval of information, especially the free-form textual information an academic must track and manipulate in order to write lectures, books, and articles. So when a new piece of software, Palimpsest, turned out to be created especially for people like me, it didn't take Nostradamus to predict I'd be intrigued. And when Palimpsest turned out to combine word-processing elements with features of cool tools like HyperCard, Storyspace (see TidBITS-095), Conc, and FreeText - and written, to top it all off, using Prograph (see TidBITS-312) - I was downright interested.
Palimpsest comes from Western Civilisation, an Australian company. It started as a private way of managing thousands of pages of legal documents; now it's released to the world for managing, investigating, and relating electronic documents generally. (A palimpsest is a manuscript that has been rubbed out and written over, and no, I didn't have to look it up; I used to be a classicist, remember?) You can learn more about Palimpsest at their Web site, or download a demo from Info-Mac.
The Basic Milieu -- You use Palimpsest to read, create, navigate, and investigate Palimpsest documents. If your documents aren't initially Palimpsest documents, you can create a Palimpsest document and either paste (or drag & drop) material into it, or import material as styled text from SimpleText.
Using Palimpsest looks and feels rather like using HyperCard. You probably will have several windows that look like HyperCard stacks, each consisting of one card dominated by a scrolling field of styled, editable text. Each "stack" is actually called a Section, and Sections are bound together behind the scenes into a Document. Each Section can itself be subdivided by Headings. Here's how Documents, Sections, and Headings are related:
A Heading is merely a piece of text to which you have applied the Make Heading command. Using a floating windoid called the Heading Browser (which displays the Headings in whatever Section is frontmost) you can give each Heading a level, so that they appear in a hierarchical, outline-like relationship to one another. (This hierarchy is purely conceptual; it has no visible analogue within the text of the Section itself.) Double-clicking a heading in the Heading Browser jumps you to that place in the Section.
Similarly, a Table of Contents window lists the Sections of the Document in a meaningful order, like chapters in a book. You can change the order by dragging the Section listings, and each Section listing can expand to show the Headings it contains. Again, you can double-click a Section or Heading listing in the Table of Contents to go there.
You can also navigate from Section to Section conveniently with a pop-up menu in the lower left corner of each Section window, which lists all Sections of the Document.
Section Types and Document Types -- Sections come in Types, which are like HyperCard backgrounds: for instance, if a typical Section of a particular Document is of the "Chapter" type, then the physical layout of each Chapter Section is identical, differing only in the contents of its fields. Documents come in Types too, each consisting of the Section types it can contain. At any time, you can alter a Document by adding a new Section of any type which that type of Document can contain.
There are also automatic Section types: the Table of Contents is itself a Section type, but you can't create a new one; every document automatically contains exactly one Table of Contents, as well as one Title Page and one Cover. You can, however, modify these Sections - for instance, you can paste a picture on the Cover.
A Document Type and the Section Types that constitute it form a template instantly affecting all Documents of that type. You can modify an existing Document type, or create a new Document type. To do so, you draw the layout of its Section types, possibly by modifying existing Section types to make the process faster. You can change the size of a Section's window (its "card" size); you can add or resize Section fields (like "card fields", their contents are unique to each Section) or Document fields (like "background fields", their contents are shared among all Sections in which they appear). You can also give a field a name, a style (e.g. scrolling or not), an initial text, initial text-style attributes, and so on. All this can be done intuitively, as in HyperCard or FileMaker.
Slicing the Cake -- What I've described thus far is a convenient method of dividing, formatting, and navigating a document, but it isn't all that different from what you might do with a word processor. The interesting part comes when you start to slice through the Document's divisions, to examine and navigate your Document in new ways.
For example, you might do a Search on a particular word or set of words you're interested in. The results appear as a new window showing every matching occurrence, one per line, each with some context around it - in effect, a customized concordance to a Document. If you double-click a line of context, you jump to that spot in the actual Document.
You might also create a hypertext link between two places in a Document. Such links are documented in a Cross-Reference Details window, showing you all links emanating from the selected text, and letting you specify a comment, an author, and a label for each link. So, you're not only linking to another location but annotating and categorizing the link as well. Later, having selected some linked passage, you can either follow the link or open the Details window. Thus, hypertext links aren't just navigational shortcuts; they're also discussions of your reasoning in associating various passages.
Palimpsest also has Annotations, which are like the comments attached to hypertext links but without linking to any other passage. Opening a passage's Annotation window is like reading a hidden footnote about it.
You can also get three sorts of "live" summaries of hypertext links (by "live" I mean you click a link to jump there):
A floating windoid called the Cross Reference Browser lists all passages in the current section from which hypertext links emanate. (Similarly, there is an Annotation Browser.)
You can obtain a list of all passages at the far end of links which emanate from the current Section or Document; this is called a Web View.
The Big, Big Picture -- So far, I've talked as if you work with only one Palimpsest Document at a time. But Palimpsest is intended to manage and relate multiple Documents. Hypertext links can run between Documents, and searches are performed over multiple Documents. What's more, there are two further entity types to help you.
First, there's the Paper. A Paper is a single window with one large scrolling text field in it, nothing more. It's meant in part as a place to take notes while you work. A Paper can also have Headings, hypertext links, and Annotations. The hypertext linking lets the Paper serve also as a repository for references to various passages in Documents, and certain special features assist with this. For instance, you can paste a passage copied from a Document into a Paper and have the pasted material be a hypertext link back to the Document passage, all in a single move. And the results of Searches, as well as Web Views and Paths, can be saved into a Paper as hypertext links, letting you quickly assemble live references to related material.
Second, there's the Study. A Study is a clickable list of Documents (and Papers) with a possible brief comment on each one. (There is also something called the Archive - there's only one - which lists all Studies.) Studies allow you to impose upon a large collection of Documents as many different categorizations as desired. Again, you can translate Search results, Web Views, and Paths into Studies to save time, and comments can help explain why you've brought these Documents together in this configuration.
So, for example, if I were writing an article on Aeschylus' Agamemnon, and I had all the scholarship on the subject over the past 40 years turned into Palimpsest Documents, then I might keep track of the scholarship in Studies - one Study listing all articles dealing with the Anger of Artemis, another listing all articles dealing with the Hymn to Zeus, and so on. Meanwhile, I could write my own article as a Paper, using hypertext links to help manage references and using the link comments to remind myself of the relationships amongst the various referenced passages.
Teething Pains -- In any software's early days there are bound to be shortcomings, and I felt there was plenty of room for small improvements in the version I saw. Some behaviors were slow. Windows didn't remember their sizes and positions. Windoids couldn't be resized, and were too small to be useful. Hypertext link labels couldn't be edited. It was difficult to know where the end of a hypertextually linked passage was, so you could easily extend it accidentally. Wider import/export capabilities (using XTND, perhaps) were needed.
However, these quibbles are minor - and temporary. Western Civilisation is a responsive company, and fully expects to incorporate fixes and user suggestions. A faster PowerPC-native version has just been released, and most of the points I raise above are slated for fixing.
Granted, Palimpsest probably doesn't do any one of its various functions as well as a program dedicated to that function alone: it doesn't process words as well as a real word processor, or manage hypertext with the ease of Eastgate's Storyspace, or build its concordances with the flexibility of Conc. The important thing, however, is that it recognizes the need to juggle, analyze, relate, read, and write about large numbers of electronic text documents. Once you've seen the Search (concordance) and hypertext tools in action, it becomes obvious how badly needed they've always been. For $50 you get a fully working copy and free updates for a year - a very decent price. If you think Palimpsest might have a place in your electronic world, you owe it to yourself to download and try the free demo.
Western Civilisation Pty. Ltd. -- +61 2 9130 1731 (Australia)
Non-profit, non-commercial publications and Web sites may reprint or link to articles if full credit is given. Others please contact us. We do not guarantee accuracy of articles. Caveat lector. Publication, product, and company names may be registered trademarks of their companies. TidBITS ISSN 1090-7017.
Previous Issue | Search TidBITS | TidBITS Home Page | Next Issue