The Operation of the Xanadu Hypertext System [current implementation 8/1982]
Retrieval of Text and Topology
The basic function of the Xanadu hypertext system is the storage, maintenance,and retrieval of arbitrarily large amounts of textual and relational information.
The Xanadu hypertext system can be thought of as a mass of text with no inherent order in itself. With the mechanisms of documents and links, any arbitrary topology can be imposed on the data that the system contains. The order that the user sees is imposed by the document. Text, link ids, and document ids are retrieved with:
- RETRIEVEV
The topology can be examined with the topology retrieval operations:
- RETRIEVEDOCVSPAN - RETRIEVEDOCVSPANSET
It is useful when dealing with a document to know how much material is involved. In the case of a very small document, a front- end may want to retrieve all of the text at once, where in a large document, only small pieces will be retrieved at any one time. This is one use for the topological retrieval operations.
RETRIEVEV <spec set> This retrieves the document ids, text, and link ids that are specified by the request. The information is returned in an order that corresponds to the order of the spec set, not in the order of the system addresses. This request can be used to retrieve the actual information that a link points to by using a link id in the place of a document id in the spec. This would be following the link, but in a more direct fashion than with the FOLLOWLINK request. RETRIEVEV will not be useful for following links when the topology of the link is of interest. For instance, to retrieve the text that the to-set of a link points to, the span in the spec would have a stream address of 2.1 and a width of 1.
RETRIEVEDOCVSPAN <doc id>
This returns the span of the document. This span is the extent of the entire document, with both the text and the links included. This means that if there are no links in the document, the width of the span will be the number of characters in the document. Otherwise, the width of the span will indicate the number of links in the document.
RETRIEVEDOCVSPANSET <doc id> * The current implementation of this request is incomplete.
This request is supposed to return two spans: the span of all the document's text, and the span of all the documents links. If there is no text or no links in the document only one span will be returned.
Creation of Documents and Versions of Documents
Definitions:
[1] documents
A document is a user created entity that contains text or other byte information and/or links. A document is divided into two distinct regions; a text space and a link space.
[2] versions
Versions are created from previously existing documents with the CREATENEWVERSION request. A version is created when you want to preserve and intact copy of a document as is, and have another copy to edit. A new editing incarnation is NOT a new version. The Xanadu hypertext definition of this term is very specific. A version is a document, the only distinction being that a version is descended from a previously existing document. A new version inherits all of its immediate ancestor's information, both textual and topological.
The Xanadu hypertext system has the following functions for the creation of documents and versions of documents:
- CREATEDOCUMENT - CREATENEWVERSION
A front-end program may let the user create documents or versions implicitly, explicitly, or both, depending on the particular front-end.
CREATENEWDOCUMENT
This request creates a new document under the account and node of the user who issued the request. The document id is the tumbler address of the new document. It will be appended to the account of the user who created it. For example, if the user's last document is 17.0.23.0.15, the new document will be 17.0.23.0.16.
CREATENEWVERSION <doc id>
This request creates a new version of document <doc id> and returns the id of the new version. The new document will have a tumbler address which indicates that the new document is descended from the original. For example, if there is only one version of document 23.0.17.0.256, this request will create a document with the id 23.0.17.0.256.1. Let's describe a more complicated case. Say that for document 23.0.17.0.256, the versions
23.0.17.0.256, 23.0.17.0.256.1, 23.0.17.0.256.1.1, 23.0.17.0.256.1.2, 23.0.17.0.256.2, and 23.0.17.0.256.3
exist. Now, if a new version of 23.0.17.0.256.2 was to be made, it would have the id 23.0.17.0.256.2.1. This will have no effect on any existing version, 23.0.17.0.256.3 will still be the same document at the same address as before. The new document will inherit ALL information that was associated with the original, but be treated as an independent document, existing simultaneously with the original or any other versions.
Comparison of Documents and Versions of Documents
The Xanadu hypertext system has the following functions for the comparison of documents and versions of documents:
- SHOWRELATIONOF2VERSIONS - FINDDOCSCONTAINING
SHOWRELATIONOF2VERSIONS <spec set> <spec set>
This request returns a list of ordered pairs of all parts of the two spec sets that correspond to the same original source.
FINDDOCSCONTAINING <spec set>
This request returns the ids of all documents that contain the same material as specified by <spec set>.
Editing of Text in Documents
The Xanadu hypertext system has the following operations for the editing of text:
- INSERT - DELETEVSPAN - REARRANGE - COPY - APPEND
INSERT <doc id> <v-address> <text>
This request inserts <text> at <v-address> in document <doc id>.
DELETE <doc id> <v-span>
This request removes the v-span from the document <doc id>.
REARRANGE <doc id> <cut set>
The <cut set> consists of two to four v-addresses within the specified document. A two cut rearrange will simply delete everything between the cuts. A three or four cut rearrange transposes the two regions of text. With three cuts, the two regions are from cut 1 to cut 2, and from cut 2 to cut 3, assuming cut 1 < cut 2 < cut 3. With four cuts, the regions are from cut 1 to cut 2, and from cut 3 to cut 4, here assuming cut 1 < cut 2 and cut 3 < cut 4.
COPY <spec set> <doc id> <v-address>
Copy causes all of the information specified by <spec set> to be virtually inserted at <v-address> in document <doc id>.
APPEND <doc id> <text>
Append inserts <text> after the last character in document <doc id>.
Links: Use, Content, Creation and Retrieval
There is one operation for the creation of links.
- MAKELINK
The Xanadu hypertext system has these operations that are involved in the retrieval of links:
- FOLLOWLINK - FINDNUMOFLINKSFROMTO - FINDLINKSFROMTO - FINDNEXTNLINKSFROMTO - RETRIEVEENDSETS
Link restrictions
There are three types of restriction presently implemented:
From-set restriction To-set restriction Home document restriction
A three-set restriction will be added once link types are implemented. The fro m-set, to-set, and three-set restrictions are spec-sets. The home document restriction is a set of doc-ranges. A link retrieval operation will only return those links that meet all of the criteria set by the restriction.
MAKELINK <home doc> <end sets>
This request creates a link with the specified end-sets in document <home doc>. The new link will be appended to the link space of document <home doc>. For example if <home doc> is 1.0.2.0.3, and the last link in it is 1.0.2.0.3.0.2.15, the new link will be 1.0.2.0.3.0.2.16.
FOLLOWLINK <link id> <which end>
This request returns a spec-set that is the end of link <link id> specified <which end> to be used later by such requests as RETRIEVEV.
FINDNUMOFLINKSFROMTO <restriction set>
This request returns the number of links that fall within the restriction set.
FINDLINKSFROMTO <restriction set>
This request returns the ids of all the links that fall within the restriction set.
FINDNEXTNLINKSFROMTO <restriction set> <link id> <n links>
This request returns the ids of the next n links after link <link id> that fall within the restriction set.
RETRIEVEENDSETS <spec set>
This request returns a spec set of all the from-sets and a spec set of all the to-sets that fall within <spec set>.