BRINGING EJOURNALS ALIVE List: Hyperjournal 11 Sep 1996 From Will G Hopkins, University of Otago, Dunedin, New Zealand I have just noticed an interesting letter to the editor in the July issue of Medicine and Science in Sports and Exercise (MSSE), which is the flagship of the sport science and medicine community and the official organ of the American College of Sports Medicine. In summary, a group who have started up the Physical Activity and Health Network ... are calling on exercise scientists to "start running in the fast lane" of the Net. The response from the editor of MSSE was that they would not allow citation in MSSE of anything that was on the Net, and that anything available on the Net for public access would not be able to be published subsequently in MSSE. The editor gives the following reasons: impermanence of Web sites, lack of availability of Web-only material to the general readership of MSSE, lack of page numbers in Web-only material, and copyright. Have any of you met this attitude with other printed journals? What can be done to overcome it? Will Will G Hopkins Physiology and Physical Education University of Otago, Dunedin, NEW ZEALAND 12 Sep 96 From Steve Hitchcock, Open Journal Project, University of Southampton Isn't this just further evidence of the incompatibility between paper and online journals? In another thread on this list (Acrobat - printer driver) Bill Ball points out: >In a transitional era many publishers and readers of print journals >would like correspondence between the print and electronic versions -- the >same "look and feel", identical pagination for citation purposes, etc. Even in a transitional period one has to ask: why? There will be paper journals and online journals but they will be different, as the term 'transitional' implies, so why not start planning now for this inevitability? To take some of the points forwarded by Will Hopkins, which are by no means unique among journal editors and publishers: >The editor gives the following reasons: impermanence of Web sites, lack of >availability of Web-only material to the general readership of MSSE, lack >of page numbers in Web-only material, and copyright. 1. Impermanence: there is a point here. The long established system of citation used by journals demands a high degree of reliability in archiving (not accessibility, BTW), that is, there will always be at least one copy somewhere in the world. In contrast Web sites can change as can the copy contained, and clearly this can't be compatible with the above objective. This will be extremely difficult to solve, and it's not obvious that it ever will be since the Web is evolving as a dynamic medium, not retreating to become a static one. This is a major problem for online journals, and it won't win them many authors because on this point the paper journals have the weight of the academic system behind them. Many online journals, not just 'parallel' online journals, will try to solve this with the 'look and feel' approach but this is too conservative. It is maintaining the convention that is important, not to 'look' conventional. 2. Lack of availability of Web material: this seems perverse even today. 3. Lack of page numbers: in a related sense this is a real problem for online journals. It will be important that online readers can locate sections of text precisely, and while metadata and search engines are one solution, they are not precise or universal enough yet to be the complete answer. The same applies to linking. For efficient and fast text retrieval we will need to be able to link to specific locations in documents without scrolling, but the Web doesn't support this. The fact that paper journals can't approach this capability should not obcure that for Web users this will become a minimum objective. 4. Copyright: I guess this refers to the point that that although you may have access to a document on the Web you do not necessarily have the right to read it under its copyright protection. There is an anomaly in the law here, and publishers would be the last to be seen to be encouraging people to break this law. However, the Association for Computing Machinery (ACM), a major learned society, has publicly stated that it will treat links as citations. While this is no big deal for anyone committed to online publishing, it is a position that offline publishers would do well to follow. Of course some people will be negative about online journals because it threatens the role they have established for themselves in the world of paper journals. Online journals have very real problems, but these should be tackled with reference to the current needs of authors and readers and of the academic system in general, not with reference to the needs of publishers or other (paper) journals. It just happens that paper journals have been remarkably successful at serving these needs, but whether they serve _current_ needs is another matter. Steve Hitchcock Open Journal project Department of Electronics and Computer Science University of Southampton SO17 1BJ, UK sh94r@ecs.soton.ac.uk 12 Sep 1996 From Ezri Carlebach, BPS Journals Office Steve Hitchcock wrote: "Of course some people will be negative about online journals because it threatens the role they have established for themselves in the world of paper journals". Of the "current needs of authors and readers and of the academic system in general" he claims "It just happens that paper journals have been remarkably successful at serving these needs". No it doesn't _just happen_! It takes the combined professionalism, skill and integrity of the authors, peer reviewers and publishers of these journals to make them so successful. As for "the incompatibility between paper and online journals" this is symptomatic of a neophiliac approach which lauds all technological development as "inevitable". As Seely Brown and Duguid point out: "Technophiles and bibliophiles have engaged in rather fruitless battles in which the former point out that you can't search or link hard copy documents, while the latter point out you can't read on-line documents in the bath or on the beach, or even at your desk with much ease." however, "creative use of new documents no longer involves direct challenges to old ones, with on-line forms replacing hard-copy predecessors. Rather, these new forms appear to reinvigorate the old, extending their useful social life not ending it." (John Seeley Brown and Paul Duguid, The Social Life of Documents, First Monday: http://www.firstmonday.dk/) At the recent LibTech conference on UK Journals on the Internet, Mr Hitchcock described a page from a scientific journal as "dead" because it contained no hypertext links. This shows a failure to understand that it is the context and the community of any document that give it life and meaning. As John Milton wrote in Areopagitica "Books are not absolutely dead things, but do contain a potency of life in them to be as active as that soul whose progeny they are." The debate needs to be reframed - not all concerns about the processes and practices of new media technologies (particularly copyright, archiving, citation etc) are based on some perceived threat to "established roles", especially those of print publishers whose journals serve the academic community so "remarkably", but rather on a desire to see those needs served as well by electronic technologies as they have been by paper. Ezri Carlebach BPS Journals Office 13a Church Lane East Finchley London N2 8DX UK BPS Journals Office 13 Sep 1996 From Steve Hitchcock Overlooking some imprecision in the way in which Ezri Carlebach quoted me, I feel I ought to respond to one point since it introduces a new element in the discussion and is important in view of the work we are doing in the Open Journal project: >At the recent LibTech conference on UK Journals on the Internet, Mr >Hitchcock described a page from a scientific journal as "dead" >because it contained no hypertext links. What I meant to refer to, and I think I did refer to, were _online_ journals, and the remark was made in the context of a quote from Ted Nelson, the man who invented the term hypertext, to the effect that an electronic document without links is a dead document. Noting that many online journals today do not have any links out to the literature I went on to say that these journals haven't yet had the chance to come alive. Hence the statement we have come to associate with the project, that we are 'bringing journals alive on the WWW' by providing link services, or the ability to add links to published documents. >This shows a failure to understand >that it is the context and the community of any document that give it life and >meaning. This is exactly the point, that the link _is_ the context and the community of online documents. All the value of publishing can be incorporated in the link. I could say a lot more on this, but fortunately it has all just been said on a parallel list, VPIEJ. For those that subscribe to that list, see the mail dated 12/9 from Peter Boyce of the American Astronomical Society. For those that don't, it appears that VPIEJ doesn't have an up-to-date Web archive of messages, so I have asked Peter to consider reposting here. >As John Milton wrote in Areopagitica >"Books are not absolutely dead things, but do contain a potency of >life in them to be as active as that soul whose progeny they are." Who am I to disagree, but I should emphasise again that I was referring to online scholarly works and not seeking to detract from books or other paper publications, even journals, in any way with this remark. Steve Hitchcock sh94r@ecs.soton.ac.uk 13 Sep 1996 From Tony Grant Steve Hitchcock wrote: > 3. Lack of page numbers: in a related sense this is a real problem for > online journals. It will be important that online readers can locate > sections of text precisely, and while metadata and search engines are one > solution, they are not precise or universal enough yet to be the complete > answer. There is no problem designing sites with the same number of pages as a paper journal. This is a design problem, not a technology problem. > The same applies to linking. For efficient and fast text retrieval > we will need to be able to link to specific locations in documents without > scrolling, but the Web doesn't support this. The fact that paper journals > can't approach this capability should not obcure that for Web users this > will become a minimum objective. Excuse me?!!! Yes it does, another design issue. The problem with these design issues, is that the funds to publish electronicaly, using the technology to its full capacity are often eaten up by publishing a paper version. If you pay someone to lay out a paper journal, you do not want to pay someone to lay out an electronic one as well. Cheers Tony Grant - electronic document designer... ____________________________________________________________________________ Tony Grant mailto:tg001@dial.oleane.com Consulting and e-document design http://www.pakeha.com ____________________________________________________________________________ 13 Sep 1996 from Steve Hitchcock I'm intrigued, Tony. If I create a link from one Web document to another, say, long document, how do I link directly to a specific word or phrase in that document without scrolling, rather than hitting the top of the other Web page? How do you do it? Steve Hitchcock sh94r@ecs.soton.ac.uk 13 Sep 96 From Tony Grant By using the tag around the word or paragraph or note or image,whatever, and linking via . HTML 1.0 compliant =;-] Example: AFRO-AMERICAINS - ALGERIE - ANGOLA - ARMENIE/UKRAINE
and so on. This can be seen in action at : . The links are from a list of countries to a series of pages, for example Asia is split into six different pages. Now if academics were to get their heads together and set up a paragraph naming convention common to all academic publications, worldwide, then you could link to any paragraph you want in any HTML document _even_ if you didn't create it! Now wouldn't that be useful for quoting someone... Pitfalls and drawbacks: For the moment you have to create the documents in order to create the links. Or when you check the source code of a quoted document, links are already in place. You could always drop an e-mail to the author you want to quote and ask him/her to put the links in for you. If the document is too long, as they often are on academic sites, you don't see the page coming up until the browser has parsed all the HTML before the paragraph you are linking to. In the example above there are documents that are 60 to 80 K and it does take some time to load them. If you make a mistake refering to the #name, you do go to the top of the page it is on, and then you have to scroll down... Clever HTML document quoting: Use Javascript to bring up another window for the quoted document. This way the document you are linking from stays open in the background and after you have read the quote, or bookmarked it, or whatever, when you close its window you are back to where you jumped off. Like I said, this is a document design problem. HTML and Javascript are very powerful tools for making interactive hypertext/hypermedia documents, and this is what the CERN people [designed] the Web for, remember... In my future sites I will be using Javascript to document all links leading out of the sites. A secondary window will come up, and the choice will be "View this link in this window" or "View this link in the main window, goodbye", and/or the destinations description will be put in the status bar. My general fee for this type of consulting is 500 French Francs an hour, but HTML coding of highly interactive HTML documents will only set you back 125 FF/hour =;-] Cheers Tony tg001@dial.oleane.com 18 Sep 1996 From Steve Hitchcock In a recent mail on links and their role in transforming the publishing process I referred list readers to a posting elsewhere by Peter Boyce of the American Astronomical Society. For those that haven't seen it, here is a slightly shortened version (posted with Peter's permission). Note that the arguments below are not hypothetical, but are being applied in real publishing. Those of us in research seek to push back the boundaries; in the case of links it is remarkable how rapidly the real world is catching up. Peter Boyce wrote: >12 Sep 1996 > >... one thing we have to keep in mind about online journals is the difference >between simply delivering electronically a copy of an otherwise paper >journal and producing a truly electronic journal; one with features >which transcend the limitations of the paper page. > >Most electronic journals today still seem to be simply using the Net to >deliver a copy of a journal which could otherwise be published on >paper. The use of the PDF format for online journals is an example. >This, to my mind, is not a true electronic journal. > >A true electronic journal will fit the landscape screen, not the >portrait orientation of a paper page. A true electronic journal will >have navigation aids -- links, a table of contents for each article, >links to sections, links both to and from figures and, likewise, links >to and from the reference page. These are internal features which make >it easy to read the article on the screen. > >However, the feature which most excites the readers of our journal >(Electronic Astrophysical Jounal Letters http://www.aas.org/ApJ/) is >the "external" links to other articles. We link to the abstracts and >scanned text of referenced articles. (In a collaboration with the >Astrophysics Data System, we now have this for about 2/3 of the >references in our journal). This is what turns a single electronic >journal into a networked information resource. > >Once such links are added, and readers see the advantage of the links >to outside sources, such links will have to become part of every online >article. At this point, the paper and electronic versions will diverge >very rapidly with time, and we will, indeed, have to be careful which >version is referenced. In astronomy, the tradition is to refer only to >the beginning of an article, so the page number question does not arise >at this time, but the problem of being able to refer to specific >sections of an article is not yet solved in general. > >Another "external" link which appeals to our readers is the list of >forward references -- i.e. articles written after the article being >read, and which refer back to the article. In our journal, each >article carries with it an updated citation index of electronic >articles which refer to it. This is a really powerful tool for the >readers, and one which completes the transformation of the electronic >journal into a broad electronic information resource, first with the >addition of papers (and links) which the author considers important and >second, with the list of papers whose authors thought enough of the >paper to cite it. > >Once you start thinking of the electronic journal in this truly >interlinked way, the electronic journal and paper journals are seen as >very diffferent entities. At that point, I maintain, the paper >journals (particularly ones with electronic counterparts) will fade >away. In astronomy, this will happen within four years. (By January >1998, 95 percent of the world's peer reviewed astronomical literature >will be available on line!) > >Paper journals will not serve as the archive, since they will neither >deliver the links to outside information in a way which can be used, >nor will they be able to provide updated reference information which >the readers will come to expect. > >Electronic journals are already part of the world's biggest library -- >the Internet. Those which recognize that fact and use it to deliver >more and better information will flourish. Those journals which, >constrained by centuries of paper tradition, do not restructure their >format and expand their content to lead readers beyond their own pages >are doomed to fade away. Much of the value of the online journal will >come to reside in the links and other ancilliary information which >will come with it. [edited] >--Peter-- >____________________________________________________________________ >Dr. Peter B. Boyce > pboyce@aas.org >Senior Associate and past Executive Officer Fax: 202-234-2560 >American Astronomical Society Ph: 202-328-2010 >____________________________________________________________________ > 19 Sep 1996 From Tony Grant Peter Boyce wrote: > >Most electronic journals today still seem to be simply using the Net to > >deliver a copy of a journal which could otherwise be published on > >paper. The use of the PDF format for online journals is an example. > >This, to my mind, is not a true electronic journal. It is a step forward... > >A true electronic journal will fit the landscape screen, not the > >portrait orientation of a paper page. A true electronic journal will > >have navigation aids -- links, a table of contents for each article, > >links to sections, links both to and from figures and, likewise, links > >to and from the reference page. These are internal features which make > >it easy to read the article on the screen. This is a job for Javascript; in Acrobat there is not (yet) a serious scripting language to make forward-backward links work like above. > >In astronomy, the tradition is to refer only to > >the beginning of an article, so the page number question does not arise > >at this time, but the problem of being able to refer to specific > >sections of an article is not yet solved in general. This could be solved by use of the tag following predetermined standards as I suggested previously. > >Electronic journals are already part of the world's biggest library -- > >the Internet. Those which recognize that fact and use it to deliver > >more and better information will flourish. Those journals which, > >constrained by centuries of paper tradition, do not restructure their > >format and expand their content to lead readers beyond their own pages > >are doomed to fade away. Much of the value of the online journal will > >come to reside in the links and other ancilliary information which > >will come with it. Hear hear!!! Cheers Tony tg001@dial.oleane.com 19 Sep 1996 From Mark Steinberger, Editor in Chief, New York Journal of Mathematics Peter [Boyce] has made excellent points regarding links within electronic journals. But the linkages _to_ electronic journals from other resources are at least as important. For instance, many journals are WAIS indexed, often together with other resources. For that matter, electronic journal pages are indexed by global search services such as Alta Vista. Queries to such indices return links to the journal articles, making the information much more accessible than it would be by any other method of delivery. Also, some of the major reviewing journals are now accessible by WWW (e.g., Mathematical Reviews and Zentralblatt, in mathematics). This permits query searches for reviews, as well as cross reference links, both between reviews, and from reviews to the actual papers, provided the latter are available online. In other words, placing both journals and the reviewing journals on the Web permits the full utility of library searching to be conducted online, at much greater speed than would be possible in a physical library. One can chase references, read reviews, and fetch the actual documents of interest. No other current delivery mechanism will permit this. --Mark ---------------------------------------------------------------------- Mark Steinberger -- mark@csc.albany.edu -- Dept. of Math. & Stat., SUNY at Albany, Albany, NY 12222 Editor in Chief, New York Journal of Mathematics http://nyjm.albany.edu:8000/nyjm.html ---------------------------------------------------------------------- 19 Sep 1996 From Tony Grant Mark Steinberger wrote: > Peter Boyce writes: > > Most electronic journals today still seem to be simply using the Net to > > deliver a copy of a journal which could otherwise be published on > > paper. The use of the PDF format for online journals is an example. > > This, to my mind, is not a true electronic journal. > > That is not true if the PDF contains extensive internal and/or > external reference links. Such links make it a true hypertext > document, but one which can also be typeset tastefully. Indeed, html, > which seems to be the base format for Peter's journal, is awkward for > material containing a large amount of mathematical notation (or other > use of graphical fonts, e.g., non-Roman characters). But such material > may easily be set in pdf. Japanese is still in beta... It has existed on the WWW for a while. On the other hand Math notation exists in .pdf and has been discussed for donkey's years in Web circles... > With pdf, one can get the best of both worlds (paper and electronic). As soon as there is a decent scripting language for links. Other shortcomings of Acrobat: - too much RAM overhead - lacks much of the flexibility provided by HTML/scripting language - we are still waiting on an indexing tool for .pdf on the WWW which works with _all_ types of WWW server hardware (very important in schools which have a lot of Mac iron serving WWW sites) For the moment we are stuck with a cocktail of technologies, which in itself may not be such a bad thing. Cheers Tony tg001@dial.oleane.com 20 Sep 1996 From Steve Hitchcock >> If I create a link from one Web document to another, >> say, long document, how do I link directly to a specific word or phrase in >> that document without scrolling, rather than hitting the top of the other >> Web page? How do you do it? Tony Grant replied: >By using the tag around the word or paragraph or >note or image,whatever, and linking via HREF="URL/file.html#your_name_here">. HTML 1.0 compliant =;-] Thanks Tony, and Jonathan Griffiths of The Royal Society of Chemistry who replied privately and gave a similar answer. I should have been more specific in posing my question, however, because I want to be able to link to specific points in documents that I don't own or didn't create, which is a limitation of this coding, as Tony pointed out. >Now if academics were to get their heads together and set up a paragraph >naming convention common to all academic publications, worldwide, then you >could link to any paragraph you want in any HTML document _even_ if you >didn't create it! Now wouldn't that be useful for quoting someone... > >Pitfalls and drawbacks: > >For the moment you have to create the documents in order to create the >links. Or when you check the source code of a quoted document, links are >already in place. You could always drop an e-mail to the author you want to >quote and ask him/her to put the links in for you. In the Open Journal project we are concerned with links which are applied by a means other than html coding. Two effects are relevant here: one, we can make many more links; two, we can provide links that point from, not just to, third-party documents. This is a system for link _publishing_ rather than simply link authoring, and in the 'open' publishing system that Web has become it will become more important to be able to add value, through links for example, without demanding ownership (copyright etc.) as publishers do now. But in this new environment of link publishing there is a need for much greater precision in linking. Whether it can be solved via Web standards or the sort of consensus that Tony suggests I don't know, but it may not be something that can be solved in isolation with a technical quick-fix. >My general fee for this type of consulting is 500 French Francs an hour, >but HTML coding of highly interactive HTML documents will only set you back >125 FF/hour =;-] Since your reply is mediated by Hyperjournal, you can bill Damien direct. Steve Hitchcock sh94r@ecs.soton.ac.uk 20 Sep 1996 From Tony Grant What do you mean by "make many more links"? > Whether it can be solved via Web standards or the sort of consensus that > Tony suggests I don't know, but it may not be something that can be solved > in isolation with a technical quick-fix. A mix of both, it seems. I like using low tech, existing technology to resolve problems. This sounds like a plea for browser editors to include more support for the tag which does a fair amount of what you desire. The (or modified version of) anchor could be used to point to an external document holding all links refering to the document, as well as the local referer (if one exists). This external document will hold the URL of links that refer to different tags in this particular document. So when you link to, then want to go back you could use scripting to put up a dialog box "Back to previous page/link" and "To other pages that refer to this link". When you go to this page of links from a document the references to specific paragraphs are contained there [Document refers to ] Say by a [key combination-click] on the first word of a paragraph you go to the page of links to the paragraph. I would, personally, put up all external links from the original document in a new window, that way to "go back" to it you just need to close the secondary window. I would prefer to see the links "from" the document being reffered to, being from the whole document. This way the new standard would need to define how people can write to this "list of links to doc X" which seen from doc X becomes "list of links to parts of this document". Instead of throwing this at one of the HTML groups it could be implemented _today_ with the help of writers and webmasters. Is Ted Nelson reading this thread? =;-] > >My general fee for this type of consulting is 500 French Francs an hour, > >but HTML coding of highly interactive HTML documents will only set you back > >125 FF/hour =;-] > > Since your reply is mediated by Hyperjournal, you can bill Damien direct. In fact I was hoping to be hired by the Open Journal Project... =;-] Cheers Tony tg001@dial.oleane.com 20 Sep 96 From Kerry L. Kresse, Physics Librarian, University of Wisconsin -- Madison >Steve Hitchcock wrote: >> In the Open Journal project we are concerned with links which are applied by >> a means other than html coding. Two effects are relevant here: one, we can >> make many more links... I, too, am concerned about Steve's meaning of "make many more links." I can appreciate the possiblities that all of these added links will provide, but I wonder who will keep track of them? With all of this talk of adding links and adding value, what happens when the site closes or moves? I understand that the creation of PURLs will catch changing addresses and urls. Will this be sufficient? In addition, creating forwards and backwards links is time-consuming work, and companies like the Institute for Scientific Information charge large sums for exactly this kind of information. I have been reading these recent exchanges about the astronomy journals with interest. I think that the astronomy community has done an excellent job of electrifying (?) their literature, and the fact that the community is small and cooperates well together has a lot to do with it. I am less convinced that broader disciplines, such as chemistry, or biochemistry, would be as successful in this kind of endeavor. I would love to be proven wrong. It could be that smaller segments of disciplines (high energy physics, organic chemistry) would be able to implement such a scheme. Interdisciplinary journals could be too much of a challenge... Aren't you glad it's friday? .................................................................. Kerry L. Kresse, Physics Librarian Physics Library University of Wisconsin -- Madison 1150 University Avenue Madison, WI USA 53706 << kresse@doit.wisc.edu >> 23 Sep 1996 From Steve Hitchcock >>What do you mean by "make many more links"? To be specific, more links than are found in most online journals that are available on the Web today. Also, I should say we are not concerned with more for the sake of more, but are also concerned with the quality and maintenance issues. >I, too, am concerned about Steve's meaning of "make many more links." I can >appreciate the possiblities that all of these added links will provide, but I >wonder who will keep track of them? A link publisher keeps track of them. As documents change and new documents are added, the link publisher updates the links. The publisher does this as a commercial service. Users will want more up-to-date quality links to navigate the research literature, and will pay for this. So sell them a subscription to a continuously updated link service. Our approach is to separate the link data from the document data (nobody wants me to go into this now, I'm sure), and the effect is to turn the linkbase into a commodity that could be sold in this way. >In addition, creating forwards and backwards links is time-consuming work, and >companies like the Institute for Scientific Information charge large sums for >exactly this kind of information. As systems develop and new link publishers emerge charges will perhaps become competitive, but there will still be a commercial charge for this type of service, and why shouldn't there be if it is what people want, in other words, if links are to become recognised as the new value-added feature of the online journal? Steve Hitchcock sh94r@ecs.soton.ac.uk 24 Sep 1996 From Tony Grant >>What do you mean by "make many more links"? Steve hitchcock wrote: > To be specific, more links than are found in most online journals that are > available on the Web today. Also, I should say we are not concerned with > more for the sake of more, but are also concerned with the quality and > maintenance issues. OK, understood. > Our approach is to > separate the link data from the document data (nobody wants me to go into > this now, I'm sure), and the effect is to turn the linkbase into a commodity > that could be sold in this way. The only service that I am aware of that does this at the present time is Alta Vista. But your page has to be indexed by Alta Vista, _and_ the sites that are linking to it do too. Usually they are, because the Alta Vista robot crawls up all links in a document that it is indexing, and then links from that doc etc., but there is a noticeable delay in updates. Human intervention is and has always been necessary for quality work, nice to see that Steve has thought about us "information scouts" in his plan =;-] This kind of information search and spying on competing business is a full- time job for many companies in the non academic WWW. So this is either an extension to existing link hunting/maintaining businesses, or new startups with an academic clientele. > >In addition, creating forwards and backwards links is time-consuming work, and > >companies like the Institute for Scientific Information charge large sums for > >exactly this kind of information. > > As systems develop and new link publishers emerge charges will perhaps > become competitive, but there will still be a commercial charge for this > type of service, and why shouldn't there be if it is what people want, in > other words, if links are to become recognised as the new value-added > feature of the online journal? Someone said "large sums of money" =:-] What is a large sum of money? How much is information worth? France Telecom says it is worth up to 3 FF a minute of connection time, even if you don't find what you are looking for... Cheers Tony tg001@dial.oleane.com 24 Sep 1996 from Steve Hitchcock Tony Grant wrote: >The only service that I am aware of that does this at the present time is >Alta Vista. But your page has to be indexed by Alta Vista,_and_ the sites >that are linking to it do too. Usually they are, because the Alta Vista >robot crawls up all links in a document that it is indexing, and then links >from that doc etc., but there is a noticeable delay in updates. Which is why we, and one or two other groups, are developing a new type of service - a link service - that is entirely different from Alta Vista. AV is good at searching and indexing, as you say, but we are concerned specifically with linking and the cognitive element associated with links, i.e the assurance that you are going to find something good and relevant at the end of a link. Again, I won't go into detail here; the proof of the pudding... , and we hope to be able to let users try it out soon. Basically we have a proven system for making links (separating the link data from the documents and storing this data in link databases); it is the user interface that needs further development so users can understand how to use this, at first glance, unusual approach. Given that we are not the only group working on such an approach, my contention is that we will at some time see the emergence of link publishers, and their function will have more in common with conventional publishing than an automated Web service such as AV. Steve Hitchcock sh94r@ecs.soton.ac.uk 25 Sep 1996 From Tony Grant Bonjour, All is becoming clear, are the different groups collaborating on a standard way of implementing this? How are you finding the pages that are quoting sections of the target document, through the log files? Your own WWW Worm? Is there a standard way of identifying parts of documents yet (other than the anchor)? I want to build this into sites as soon as possible. I am also interested in how you are identifying the parts of documents that are targets of links from other documents. I am not completely satisfied with the Icons used by John December, I am trying to find something more "general public". Cheers Tony tg001@dial.oleane.com 27 Sep 1996 From Steve Hitchcock Tony grant wrote: >All is becoming clear, are the different groups collaborating on a standard >way of implementing this? There is contact at a research level, of course, but so far only tentative steps towards any form of standardisation. It is very important here to distinguish between the technical issues and the publishing framework. Standards can be important in providing access to technology, but publishing itself has never been a standardised activity. What these systems have in common is that they are 'open hypertext', in a technical sense, meaning that in applying hypertext links they are not tied to specific or proprietary document formats. There are different degrees of 'openness', however, and this relates to ways in which the documents and link databases (these are stored separately, remember) are managed. In the Open Journal project we are using the Distributed Link Service, developed here at Southampton, which at the moment applies no control over the management of documents but is concerned with the management of links. In this sense it is consistent with the 'open' publishing philosophy of the Web. Another link service, Hyper-G (or HyperWave) developed in Austria, is 'partially open' in that it uses an 'open' link service as above but it demands some degree of control over the document management, i.e. a document has to be registered, or deposited, on a Hyper-G server for the link service to be applied. Hyper-G has been compared with Ted Nelson's Xanadu project, which began back in the 1960s. Also, Hyper-G predates the Web, and on occasion its developers still argue not that it complements the Web but that it will supersede it. There are technical merits on both sides, and these can be argued over. Standards. collaboration, openness are important from the usability point-of-view, but differentiation is also important for publishing applications. Xanadu was a visionary system and has always had problems matching technology to its vision, but if the original vision is ever realised it would surely be discovered that a publishing system designed to be so complete would not find universal support from publishers. Publishers may seek a technical framework for new ways of publishing, but they don't want one where all the publishing rules are defined for them as well. It has been put to me that the various projects like ours should get together and work on one single development - strength in numbers, viability and so on, it is argued - but as I've tried to show, this is not necessarily the best solution. Yes, standards, up to a point, and collaboration; but differentiation and competition are also important motivators in publishing, which ultimately is what we are all concerned with. >How are you finding the pages that are quoting sections of the target >document, through the log files? Your own WWW Worm? A combination, including iterative and automated database searches and, for the future, intelligent agents, as we now like to call them. In some cases we don't find the quoting pages at all, the reader does; these pages can still point to a target document, using what we call a 'generic' link. For example, say I had written or knew of the existence of an (online and accessible) article which was the last word in, say, biology (not likely, I know, but this is a very general example). So I will create a generic link for the word 'biology' pointing to this document. Then if someone is viewing documents, any documents, over the Web through my link service (set the browser proxy), they will find that they can click on any instance of 'biology' and they will see the document. This highlights two features of a link service application: 1. By applying the principles of 'openness' and separating link data etc., we can link FROM third-party documents even if those documents have no authored links (without violating the integrity of the document data); without having prior knowledge of their existence, these documents become linked into the literature. 2. We are always pointing TO known documents. There is a shortish piece elaborating a little on this approach in the latest Ariadne, the eLib magazine, at http://www.ukoln.ac.uk/ariadne/issue5/open/ >Is there a standard way of identifying parts of documents yet (other than >the anchor)? I want to build this into sites as soon as possible. I think this is where we came in on this thread. >I am also interested in how you are identifying the parts of documents that >are targets of links from other documents. I am not completely satisfied >with the Icons used by John December, I am trying to find something more >"general public". I shall have to find out about December's Icons. Steve Hitchcock sh94r@ecs.soton.ac.uk 28 Sep 1996 From Tony Grant > Standards can be important in providing access to technology, but publishing > itself has never been a standardised activity. Agreed, the more variety in document format the better, but with standard access if the system(s) is(are) to succeed. > Xanadu was a visionary system and has always had problems > matching technology to its vision... I like the past tense applied to Xanadu... My thoughts exactly. > Yes, standards, up to a point, and collaboration; but > differentiation and competition are also important motivators in publishing, > which ultimately is what we are all concerned with. For any document type, it is the content which is of primary importance. Being able to have easy and powerful hypertext access to a document (or parts of a document) is a plus, but one that we have lived without until now. So any system providing links to documents must be: - as powerful as possible - as unobtrusive and easy to use as possible - and, if there are competing systems, as "standard" in interface and use as possible The end user is king... > know, but this is a very general example). So I will create a generic link > for the word 'biology' pointing to this document... I think that the day when I can click on any instance of a word in, say, my browser and go to a list of links concerning that word is too far away. I want to be able to create hypertexts with to/from links "next week" with existing technology. This is "only" a standard document part naming convention away (in the present case with HTML, txt, .pdf...). That projects such as yours promise better things for the future is wonderful, but my goal is to get something working "tomorrow". What we will need later is a smooth, "invisible" transfer from one technology to another. But the WWW should already be working as it was intended to work by its creators _despite_ the size it is now. Academic publications should be leading the way in implementing useful, usable hypertext. Commercial interests are only interested in reproducing traditional publication on a new medium, very few are doing real hypertext. Cheers Tony tg001@dial.oleane.com 2 Oct 1996 from Steve Hitchcock >That projects such as yours promise better things for the future is >wonderful, but my goal is to get something working "tomorrow". We would show you tomorrow if you could come to Southampton. Since that probably isn't very practical we are trying provide an interface so it can be tested over the Web, but making it easy to use is taking a little longer. Bear with us, but it will be faster than you suggest. We are almost halfway through a three-year research project, with a number of publishers and a lot of people to satisfy. >What we will >need later is a smooth, "invisible" transfer from one technology to >another. But the WWW should already be working as it was intended to work >by its creators _despite_ the size it is now. Our system works with the Web, so in this case the 'transfer' will certainly be invisible. >Academic publications should >be leading the way in implementing useful, usable hypertext. Commercial >interests are only interested in reproducing traditional publication on a >new medium, very few are doing real hypertext. Right, and the implications are much greater than link-hopping from one page to another. Steve Hitchcock sh94r@ecs.soton.ac.uk