SCSI Confusions — Don Norman of Apple writes:

At the last Computer Bowl contest, the question of "how many SCSI IDs" was asked. One of the contestants said "eight," but this was ruled wrong by the judges who said "seven." The audience yelled. I myself called out, "number 0 is one of the ports." The question was turned over to the referees, and after consultation, they very carefully, in a measured tone of voice, announced, "The highest number is 7." This is, of course, correct, so the answer of "eight" was ruled wrong. Poor show by all involved.

Communicate Lite may replace some of the abysmal programs currently bundled with modems. The communications program from Mark/Space Softworks uses a document-oriented approach along with support for Apple’s Communications Toolbox, which allows users to add power by adding tools. Communicate Lite is available for bundling, and costs $49.95 for single copies direct from Mark/Space. A more-powerful version due this summer, Communicate, will add more communications tools, Apple events scripting for Frontier and AppleScript, automated virus detection, and integrated In and Out boxes that simplify file transfer.

Mark/Space Softworks — 408/982-9781 — 408/982-9780 (fax) — [email protected]

ZipIt Wires — Jacob Ahlqvist <[email protected]> writes:

In TidBITS #182, Jim Wheelis, in his review of ZipIt, failed to mention one great advantage of ZipIt – it is Apple event-aware and ties in completely with Kem Tekinay’s Freddie 1.2.5 (and only 1.2.5) to provide automated decompression/opening/reading of PC .QWK files downloaded from a PC BBS for off-line reading. Now you can just drop the .QWK file onto Freddie and leave it for a few minutes to do the trick rather than unzipping manually with UnZip or StuffIt Deluxe.

ClarisWorks has expanded to the Windows market, with Claris announcing that IBM and Toshiba will bundle ClarisWorks for Windows with certain computer models. If you have to buy ClarisWorks for Windows (a good way to achieve cross-platform compatibility), you can buy it for $99 until 15-Aug-93.

Claris — 800/3CLARIS — 408/727-8227

Wolf Creek Technologies recently slashed the price on QM-PAGE, their alphanumeric pager gateway for QuickMail, dropping it to $995 for 20 users and adding 10 and 5 user packs for $595 and $325, respectively.

Wolf Creek Technologies — 407/334-0448 — 407/334-2303 (fax)

A while back (November ’92, in TidBITS #152, to be exact) I said some positive things and some negative things about Now Utilities 4.0.1. Now I’d like to take back a substantial portion of the negative things. I have three reasons why:

• I complained that despite the purported fixes between 4.0 and 4.0.1, Super Boomerang and NowMenus together still caused some crashes on my machine when the Standard File Dialog tried to come up. (Since this usually happened when I tried to save a document for the first time, I wasn’t too pleased.) However, it has been a long time since my Macintosh has had one of these crashes. I think rebuilding my desktop may have helped; also, turning off Keep Show Info and Remove Unmounted in Super Boomerang’s Control Panel may have had something to do with the change. Whatever the reason, I now regard the pair as more stable than I used to.
• I said that "the lists [in the pull-down menus] are not hierarchical. Documents can be attached as submenus to programs; but programs themselves cannot be made submenus to anything. So if you want a really extensive list of your programs, you get a huge scrolling menu." A reader wrote in immediately to inform me that this was false. You can include a folder in a pull-down menu, and this folder can contain aliases of programs. If you make a bunch of folders representing categories (Font, Word Process, etc.) and put aliases of appropriate programs inside them, you can get a hierarchical arrangement of your programs by category in a menu. Furthermore, by some miracle I don’t understand, recently used documents appropriate to those programs automatically attach themselves to the names of their aliases in the menu. So much for accuracy in my reviews; anyhow, that objection is transmuted to delight.
• I lamented that DroppleMenu, one of my favorite extensions, didn’t work under NowMenus 4.0.1. That has changed, thanks to David Winterburn, whose latest version of Menu Dropper (7.1b6 is the one I saw) does work with NowMenus. This means you can drag an icon onto the Apple menu and right down through its hierarchical menus, and have the thing the icon represents be moved, copied, or aliased to a folder, or opened by an application. This puts Menu Dropper up where DroppleMenu was when I was able to use it, in my Top Ten category of extensions, namely, "Things so valuable and obvious you can’t believe Apple didn’t build them into the System (or Finder) in the first place."

Since I already said in the earlier review that no one should be without Super Boomerang, and that the whole price of Now Utilities was worth it for just this one application, it remains only to say that those of you who still haven’t bought it should (a) download and read the review, mentally correcting it to take account of these retractions; (b) try out the demo version available at most FTP sites right now; and then (c) run, don’t walk, to your phone or software store and purchase Now Utilities 4.0.1 immediately. Better still, skip (a) and (b).

Those of you who read TidBITS or other setext files on Unix boxes may wish to check out a prototype setext viewer now posted at <sumex-aim.stanford.edu> for anonymous FTP as

info-mac/text/setext-viewer-02-unix.txt

This 13K program, sv-02, requires System 5 and the curses library. It will not run on a Macintosh, except perhaps under A/UX. Oguz Isikli, a graduate student at Bilkent University in Turkey, wrote the code. Oguz ported the parsing engine from Akif Eyler’s Easy View, and Oguz based the user interface on the Unix Gopher client.

Oguz and Akif do not consider sv-02 complete, and are looking for comments, suggestions, and possibly source code contributions to the project.

Although anyone who knows enough about how to compile a program under Unix probably knows how to deal with uudecode, uncompress, tar, and so on, Akif provided the following instructions for defunking the file. Since the sumex moderators changed the name, you may wish to rename it to sv-02.tar.Z.uue before starting out.

Instructions for defunking:

uudecode sv-02.tar.Z.uue
uncompress sv-02.tar.Z
tar -xvf sv-02.tar
cd sv-0.2
make
sv tidbits-184.etx

— Information from:
Akif Eyler — [email protected]

Although Hewlett-Packard does not recommend refilling their disposable DeskJet/DeskWriter cartridges, there is little risk and much profit in refilling your own model 51608A or 51626A print cartridges. The HP DeskJet/DeskWriter color cartridge 51625 is not refillable. HP made this happen by not putting vent holes in the top of the color cartridge. Over the course of three years experimentation, the following helpful hints should save pain for the neophyte user who plans to refill cartridges.

Use only water-based ink, as alcohol-based inks will immediately dry up and clog the print head. Previously, the Parker "Super Quink" brand was the ink of choice for veteran cartridge refillers, but Parker no longer produces it. The remaining Parker ink, known simply as "Quink," will not provide good results. The best current brand for refills is known as "Skrip." Found in a yellow box, this is an extremely common brand of office ink generally used for stamp pads and fountain pens. You’ll get the best results from the permanent black ink, but colored inks such as blue and red work fine too. When experimenting with a brand, keep in mind that price is no indicator of performance.

You can refill the cartridges using an ordinary 3 cc diabetic syringe, available at most pharmacies for less than 50 cents. (In North Carolina, U.S., no permits or prescriptions are required to purchase these syringes, however that may not be true in other states or countries.) Wash out the needle with warm tap water and you can reuse it almost indefinitely. Start-up costs for your homemade refill kit should run less than $3: about $2 for the ink and less than $1 for the syringe. You can get 11 refills from a single 2-3 ounce bottle of ink at a cost of less than 20 cents per refill!

Don’t refill a cartridge that has sat empty. Refill cartridges immediately after they run out of ink. It does no good to wait and collect used cartridges because remnants of the original ink will dry up and render the unit worthless.

To refill the cartridge, assemble the ink container, cartridge, syringe and a few absorbent paper towels on your work surface. Do your work on a glass-topped or ceramic surface which will not absorb any ink spills, and be sure to place a paper towel under the cartridge. Don’t pour the ink into the top of the syringe; instead, immerse the tip into the inkwell and withdraw the plunger, sucking the ink into the cylinder. Plunge the syringe into the top of the cartridge, through the vent hole for the entire length of the needle and slowly press the plunger. If you see ink bubbles around the hole, don’t let them pop, as the ink will splash everywhere. Hold a clean paper towel around the vent hole to catch the bubbles. One cartridge will hold two injections of ink from a 3 cc syringe.

Wipe off the print head. You should have some leakage from the refill process, which indicates a successful refill. If you see no leakage, the print head may be clogged. It’s possible to unclog the head by blowing into the vent hole (carefully!) or wiping off the print head with a wet paper towel. You know the refill worked if you can wipe the print head with a tissue or paper towel and get a thick band of ink on the paper.

Using the right kind of printer paper with your refilled cartridges will provide better-than-new results. Use a paper with high cotton content and a tight fiber "weave." Hammermill Bond, Hammermill Laser Copy, St. Croix Laser/Xerographic, and Xerox 4240 provide great looking printouts from any cartridges, refilled or not. Refilled units also work fine with transparencies.

You can refill the HP 51608A up to ten times before the electrical contacts on the cartridge begin to deteriorate. I have used refilled cartridges in the same DeskWriter for the past three years with no damage to the printer.

Refilled and new cartridges should be good for about 500 pages of printing text or light graphics. If your printer gives less than 200 pages from a cartridge, there is an upgrade kit available only to early purchasers of the DeskJet and DeskWriter that greatly extends the print life of any cartridge, refilled or new. This kit is available free for affected users from Hewlett-Packard by calling 800/538-8787. I don’t have the serial number range handy.

The upgrade kit is better described as a hardware patch. The early DeskJets and DeskWriters used a cartridge cradle that wasn’t perfectly airtight. As a result, the cartridge could prematurely dry out. The classic symptom of the problem is when someone gets only 200-300 pages out of a cartridge instead of the 500 page design limit. Usually, the print quality will be terrible for most of those 200-300 pages.

The upgrade kit consists of a new cartridge cradle and a little tool with which to install it. You use the tool to remove the old parts and to install the new parts. After you install the kit, (it takes about five minutes), you will notice better print quality and many more pages per cartridge. Refilled cartridges especially like the new sled.

Hewlett-Packard — 800/538-8787

When the seller of goods is no longer a village craftsman dealing with friends and neighbours on a one-to-one basis, but a multinational company with hundreds of products and millions of end users, it is impossible to deal with each customer individually. Thus, the late industrial civilization created complex patterns for the distribution of goods and services. A manufacturer sold to a limited number of regional distributors, who in turn resold in smaller bulk lots to local distributors, who moved product to retailers in case lots, who then sold to the end user in one-of quantities.

The advantage of the distribution pyramid is its simplicity at each stage. No one level creates an unmanageable number of customer records. The disadvantage is that the price may increase by three or four hundred percent by the time an item reaches an end user – this without any value being added to the product along the way.

Already, many home-based businesses are built on short-circuiting this process. They offer soap, jewelry, clothing, cookware, and other goods directly from the manufacturer to the consumer through in-home sales representatives. However, these schemes can still be improved, for most still have distribution chains, and only the physical overhead is really saved.

Information technologies such as automated ordering/billing and computer assisted manufacturing (CAM) enable a better way. Customers could view sample goods online or in a local showroom licensed by a manufacturer and/or order items to personal specifications from a catalog. The goods would then be made to order on demand by automated assembly lines receiving computerized instructions for each item.

Electronic ordering and funds transfer would enable manufacturers to deal directly with the millions of customers. No paperwork would have to be handled, for none is created, and building to order cuts inventory and reduces costs further. This method might be most fully applied to goods requiring customizing – clothing, automobiles, and computer hardware. There is less to gain in the production of general hardware, household items and tools, for they can be identically mass produced. However, all would benefit from the shortening of the distribution chain.

There is nothing new in these ideas; indeed, they could be regarded as obvious extrapolations of current methods of doing business. Direct distribution coupled with automated ordering, manufacturing, and paperless payment is just a natural outgrowth of information age technology applied back to the problems of the industrial age. Such a development would contribute to making the industrial infrastructure as invisible as is the agricultural infrastructure today. How many people do you know who make their living growing food? More to the point, how many First Civilization people do you know – those making a living as hunter-gatherers? About as many as your children will know of factory workers and store clerks.

If this is not so far revolutionary, then how will information age techniques create new distribution paradigms? How will information and the (software) tools needed to create, manipulate, and access it be distributed and accounted for? After all, the number of contributors to a particular data bank or manipulation tool may be legion. In an age of reusable software components, the intellectual creations of scores or hundreds of people may be employed for a single information transaction. The industrial paradigm was that such techniques were licensed or purchased outright by a manufacturer, and the cost spread out over the number of items. If the new product was a success, not only did no further payment go to the creator of the enabling techniques, but the law allowed the new owner of the technique to restrict its use in other products. This may be an acceptable stopgap for hard goods in a society that is limited even in its ability to record the sales of goods, much less the use of methods, but it is already feasible to propose much better.

Define a civilization’s "metalibrary" to be the set of all its knowledge, (information and technique) together with the means of storing and accessing it. "The Metalibrary" is the universal information store, including data, journals, magazines, newspapers, books, TV programs, movies, artwork, in short, everything there is to know on whatever media. The Metalibrary already exists, but it will grow and develop to become something much more complete.

Assume that anything could be posted or read (for a fee.) Assume that all will be hyper-indexed in space and time, so that any kind of multi-media thread can be followed through the Metalibrary. Indexing threads could be attached by individuals or by editors, and a user would be free to accept for view-use any thread collections, or only those of certain editors. (Journals would become collections of threads by the responsible editor.) Every home and business would have Metalibrary Terminals of various kinds. Some would do data searches, some show publications such as National Geographic in full colour; large ones might display artwork or symphonies.

Each individual would have an indexing profile, started manually, but maintained by a "world view daemon" that monitored usage preferences. Every Metalibrary item (even the world views of others) would have a registered UIC (Universal Information Code.) This would be an index to the registry of contributors to that item, with their percentage share in the proceeds of its use. The registry would be hierarchical; one UIC might refer (with percentages) to other UICs through many levels to individual accounts.

One distribution technique would be to download every instance of an item on a rental basis from the Metalibrary store – the ultimate in centralization . In this system, no goods are sold; everything is paid for each time a local instance is created and used. Such a method by itself has the advantage of allowing for proper automatic credit to the contributor, but the disadvantage of requiring communication bandwidths that may not be feasible.

A better (self-auditing) payment mechanism was suggested by Brad Cox (Journal of Object-oriented Programming; Jun-92; Dr. Dobb’s Journal; Oct-92) in his case for reusable software components. His technique is here adapted to the entire range of Metalibrary services.

Local devices would have smart hardware accessed by distribution code contained in every software product. This code keeps a record by UIC of use instances (not purchase) whether items are copied from the Metalibrary store directly, or obtained in some other way. Periodic reports would be sent to an accounting daemon, which would employ the UIC registry to debit user accounts and credit creator accounts appropriately. The accounting code would also have to check periodically to ensure that its results had been sent and properly received, and refuse the application permission to run otherwise. This technique could be applied to the components of access or production software, as well as to the components of the data being viewed or manipulated, for all would have a UIC. It has the same access and payment advantages as the one above, and could be used in the same way, but information or tools would only have to be acquired once (saving much network bandwidth).

Software would record each use of itself and of the information it accesses (publications on any medium, the display of artwork, 3-D artistic performances, and searched data). The software creators and data generators would receive royalties automatically in proportion to their contribution to the collection. If a user synthesized new tools or data from old, UIC codes for each component would be sent to accounting with appropriate percentages. (If the new tool were made a public product, some verification of the relative value of the parts to the whole would be necessary before actually registering a new UIC code.) For instance, if the user synthesized Walter Cronkite, Marilyn Monroe, and Elvis Presley as the evening news anchors, their estates would get a cut along with the reporters who produced the items, the team that edited that particular news thread for the day, and those who created the data files for the syntheses.

The making of a hard copy (where appropriate) would fetch another premium. Artwork (wallpaper?) displayed on the condo walls would generate a time-based fee to the parent museum. Symphonic, athletic, and other performances would generate royalties according to a formula agreed to by the participants and registered for the event along with the UIC. A percentage of each transaction would finance the Metalibrary itself. The hardest distribution problems to solve, as Cox has pointed out, would be tamper proofing the hardware and the software to store and transmit accounting information. A harder problem would be the initial cross indexing and storage of all available knowledge in every form. This task could be automated for new materials, whose primary medium of publication would be the Metalibrary.

Though specific data searches might be done online to the Metalibrary (more likely a series of networked nodes than a central location), larger and less timely collections of data, artwork, performances, books, and software tools might be better distributed "hard," i.e. on a medium such as a CD-ROM or its future equivalent, the 3-D data cube. Indeed, some things might have to be distributed "hard" to prevent piracy, as there is probably no effective copy protection means in software. There would be no reason to charge for such distribution except for the out-of-pocket cost, as pay-by-use would cover the important fees.

Metalibrary Terminals would take the place of the mail carrier, telephone, TV, book reader, journal and news reader, library card, computer, and personal data assistant. Equipped at a some stage with more sophisticated interfaces, they might eventually be known as "pocket brains," though there would be no need to suppose them to be artificially intelligent. The Metalibrary would also enable the creation of personal services partnerships or "metapersons" – like present day corporations, but of limited duration and changeable structure. These would be the primary vehicle for the assembly and sale of professional services such as education, training, counselling, accounting, writing, software production, and legal services. In no case would there be a distribution chain, for all consumers would directly access providers.

A caveat: As in other cases, information technology only enables the scenario painted here. Besides the unifying aspects of these potentially global information paradigms, other forces are at work to fragment nations, stir up old hatreds, and prevent the free flow of information. Other paradigms may replace this one before it is realized. Thus, although the present distribution chain is already obsolete and overdue for replacement, subsuming its functions in the Metalibrary is only possible; it is not inevitable.

[We welcome discussion of Rick’s ideas, particularly in relation to software distribution, online in appropriate discussion groups, most notably the Info-Mac Digest <[email protected]> and on CompuServe in the TidBITS section (#5) of MACDVEN. -Adam]