dick@ccb.ucsf.edu (Dick Karpinski) (11/02/87)
In article <21526@ucbvax.BERKELEY.EDU> oster@dewey.soe.berkeley.edu.UUCP (David Phillip Oster) writes: >... any document can be annottated by anyone at anytime. >... choose to turn off any subset >... editted by editors they trust, >... each document has fast links to every document it references >and each document that references it, >flaky theories get their rebuttals attached strongly and quickly. > >How can we evolve usenet news into such a system? Note that such a system has many useful purposes: a. readers are active participants b. kudoes and flames accumulate c. specific meanings of words and phrases emerge d. each reader is a moderator available to future readers e. the documents and system can evolve gradually f. multi dimensional searching facilities develop naturally g. authors can retract (but not erase) their former work h. optical disks will have something to store etc. To answer the question: only slowly and with cooperation on many fronts by many participants. Ted Nelson & his colleague could assist with their hypertext system or its ideas. The fundamental truth is that each word will have many more readers than writers. Thus we should attend to ease of reading above all. The first useful thing I see to do is to make a version of a popular news reading program which can follow references forward and back. The first major problem is how to expire unwanted bulk text; LRU anyone? Dick Dick Karpinski Manager of Minicomputer Services, UCSF Computer Center UUCP: ...!ucbvax!ucsfcgl!cca.ucsf!dick (415) 476-4529 (11-7) BITNET: dick@ucsfcca or dick@ucsfvm Compuserve: 70215,1277 USPS: U-76 UCSF, San Francisco, CA 94143-0704 Telemail: RKarpinski Domain: dick@cca.ucsf.edu Home (415) 658-6803 Ans 658-3797
bzs@BU-CS.BU.EDU (Barry Shein) (11/03/87)
Fascinating and wonderful idea. I think the first thing to do may not be to really provide all the software etc to deal with hypertext but perhaps to decide what information needs to be stored and how and to provide the ability to ignore it until the systems are fully built. For example it would be nice if I could just enter a Hypertext reference \(ref 1043) right into my text (probably by program) and know that most USENET reader programs \(ref 4a.37) would just ignore it by some simple rule (elide or interpret backslash followed by matching parens, for example.) Given a suitably powerful escape like that experimentation should be able grow \(ref bzs@bu-cs:1477 (graphic 106.CFD)) unfettered. Anyone want to propose a suitable escape? \(link 777 (file "hypernet")) Suggestions? \(active "survey" label "hyper-escape" in-file "hquery") -Barry Shein, Boston University
tanner@ki4pv.UUCP (11/03/87)
I would suggest, for those actively interested in such things, that you select an escape sequence that doesn't look a whole lot like that used in popular WP software. Good examples of what not to select: \(xx for many "xx", is nroff/troff special char @foo for many foo, is scribe-form crud 0x80+n for most <n>, the high bit may get lost In general, I suggest that people implementing new software not implement systems which are likely to choke on many of the files existing on their systems. Sensitivity to nroff/troff/pic input, for instance, would be inapropriate in a system dealing with text which might reasonably expect to be fed such text. Sensitivity to the high-order bit in a char would be inapropriate in a text processing package which might get ported to EBCDIC environments. Tanner Andrews, Systems CompuData, Inc. DeLand
bzs@BU-CS.BU.EDU (Barry Shein) (11/03/87)
Tanner Andrews points out the importance of finding an escape for hypertext information which does not coincide with any commonly used software and some other points. Totally agreed. The best way to do this is to keep suggesting possibilities until no one in a large and diverse group can find any objection. Here's another possibility very similar to my last: \[anything] Does anyone know of any objection to this format (backslash followed by balanced square brackets) other than the slight possibility someone might want to send such a thing anyhow for no major reason (ie. not as necessary syntax for a major package.) One other safeguard could be a header field which indicates whether or not this message should be interpreted at all, as in: Format: hypertext which could be yet-another-catchall hook which is easily implemented. I do agree tho that even such a hook cannot solve all problems as obviously one might wish to send troff or TeX text hypertex'd (in fact perhaps the whole backslash idea should be canned due to it's prevalence in word-processing, maybe replace it with colon?) Just thinking out loud. -Barry Shein, Boston University
dick@ccb.ucsf.edu (Dick Karpinski) (11/05/87)
In article <8711022250.AA01367@bu-cs.BU.EDU> bzs@BU-CS.BU.EDU (Barry Shein) writes: > >I think the first thing to do may not be to really provide all the >software etc to deal with hypertext but perhaps to decide what >information needs to be stored and how and to provide the ability to >ignore it until the systems are fully built. Don't wait for some fully built system. Just make an initial stab at some useful aspect and pass it around for trial and change. >For example it would be nice if I could just enter a Hypertext >reference \(ref 1043) right into my text (probably by program) and >know that most USENET reader programs \(ref 4a.37) would just ignore >it by some simple rule (elide or interpret backslash followed by >matching parens, for example.) Let's pick some nice form for references and find out where and why it runs into difficulties. One nice way is to identify the form of references in each document by some device like /*hyperref:{:}:*/ where the "hyperref:...:...:" is the fixed part. Any characters on the outside may be used to hide the specifier from ignorant systems. The characters between the colons are then the delimiters in use for this document (or until another hyperref). This makes it easy to revise any such document to accomodate local requirements without changing its meaning. >Given a suitably powerful escape like that experimentation should >be able grow \(ref bzs@bu-cs:1477 (graphic 106.CFD)) unfettered. You have got the idea I had in mind. Who is familiar with the philosophy that WEB uses for this sort of thing?? Who (besides Larry Wall) could implement a useful article reference in rn like this?? Should the definition be more robust?? A trailing word (ferrepyh?) could avoid false hits and allow even the ":" to vary. How should the reference be indicated to the reader? How should the reader ask to see the reference? How should references to this document be found from here? How much of the HyperCard semantics should be emulated? Who cares? > -Barry Shein, Boston University Dick Dick Karpinski Manager of Minicomputer Services, UCSF Computer Center UUCP: ...!ucbvax!ucsfcgl!cca.ucsf!dick (415) 476-4529 (11-7) BITNET: dick@ucsfcca or dick@ucsfvm Compuserve: 70215,1277 USPS: U-76 UCSF, San Francisco, CA 94143-0704 Telemail: RKarpinski Domain: dick@cca.ucsf.edu Home (415) 658-6803 Ans 658-3797
glg@sfsup.UUCP (11/12/87)
In article <8711022250.AA01367@bu-cs.BU.EDU> bzs@BU-CS.BU.EDU (Barry Shein) writes: >Fascinating and wonderful idea. I agree, when can we have it available ;-) >I think the first thing to do may not be to really provide all the >software etc to deal with hypertext but perhaps to decide what >information needs to be stored and how and to provide the ability to >ignore it until the systems are fully built. Even before this we need to know what it should do and how it will be used. You may argue that we can't know that until there is some experience using systems of this type, but experience on a partially implemented, possibly hard to use system will not really help either. We need to prototype, but not just mock-ups, a system that will be used seriously by many people. If the groundwork is solid, it can evolve from there. [ suggestions about specifying a standard that can start now to enter references into messages ] >Anyone want to propose a suitable escape? \(link 777 (file "hypernet")) I think this is premature. Entering references in netnews articles does not really address what is necessary to create and build references in a hypertext system. The essence of hypertext involves evolving textual structures, you need to be able to add to items within the data base (new references, etc.), not just to append articles with back references. >Suggestions? \(active "survey" label "hyper-escape" in-file "hquery") I suggest that it is time to start working toward a fully developed distributed hypertext system. I got Computer Lib/Dream Machines back in '77 not long after I learned of the existence of computers, but I am not aware of any practical work that has been done on the concept. Is there any? In any case, I am interested in contributing my efforts to bringing it into reality. If there is not already a large body of work, I volunteer myself to lead/organize the effort. If a significant effort already exists, I am interested in making a contribution. Here are my suggestion on some possible actions to get an effort rolling: 1) Find out what has been done. Is there software that might be adapted to the UNIX/USNET environment (this seems like a good starting point, not a limiting consideration). 2) Write a draft paper outlining the capabilities and requirments for a hypertext system. Includes the beginings of a phased plan to bring it into reality, and an exploration of ways to organize, finance, etc. the software developement. 3) Present the paper at USENIX and other converences to build widespread support and interest in the concept. 4) Start a mailing list for people who want to contribute to this effort. I say mailing list because I think it should only be exposed to those willing to contribute their efforts before it is well developed. 5) Add your own . . . I have some idea of how to approach this, but I welcome any input. If no one else volunteers to pick up the ball and run with it, I will begin working on a draft paper soon, and send it to those interested for comments (the mailing list mentioned would probably be a good place to distribute it). Send me mail at attunix!sfsup!glg if you have suggestions, pointers to existing work, or even if you think I'm crazy. Gerry Gleason
einstein%bucsb.ARPA@bu.EDU (David K. Fickes) (11/13/87)
I think you folks are crazy.... lets not try to attempt a escape sequence hypertext... esp considering that much of the net is garbage... but ..if you're willing to draft a document.. I'm all for it... the volume of news is such that I doubt that it will ever be possible to archive it all in a use able format.. but I'm willing to help you try... strengths here are database design and tech writing.. Additionally, I would appreciate you sending me any material or pointers to and material that you find that may seem appropriate... We are considering a hypertext situation, only that we are using Albert Einstein's correspondence as a textual base. - thanks david -- ============================================================================== David K. Fickes Center for Einstein Studies/Einstein Papers Project UUCP: ...harvard!bu-cs!bucsb!einstein Boston University BITNET: oth932@bostonu 745 Commonwealth Avenue PHONE: (617) 353-9249 (617) 277-9741 Boston, MA 02215
mendel@utcsri.UUCP (11/17/87)
Having just returned from Hypertext '87 in North Carolina, I was toying with the idea of hypertext facilities for Usenet (HyperNet?). I was thus pleasantly surprised to see several references to this in this newsgroup, but I could not find the article that originated all the comments. I would appreciate receiving a copy of the article from the author, and I would like to participate in any communal effort that develops. -- Alberto Mendelzon utcsri!mendel mendel@csri.toronto.edu
bzs@BU-CS.BU.EDU (Barry Shein) (11/20/87)
Looking through old messages the conversation seems to have started with one from David Philip Oster, reference <21526@ucbvax.berkeley.edu>, here's the key excerpts that touched it all off: >I've been spending a lot of time listening to Eric Drexler recently, and >he has been talking a lot about: ... >He feels the answer to the bogosity question is: hypertext authoring tools. ... >Conclusion: >How can we evolve usenet news into such a system? -B