User's Guide - 2.6 Collection Types

Suche

Home

Einstellungen

Anmelden

Hilfe

2.6 Collection Types

You have read about how to insert collections on page 57. Here is a more detailed description of the different type of predefined container classes on Hyperwave Information Server and how you can use their special features to make your information management tasks even more convenient.

There are five types of collections: Collection, Language Cluster, Sequence, Multi Cluster and Alternative Cluster.

ICONS

The icons used for these collection types are listed in the Icon Guide on page 2.

COLLECTION

Collections are used to hold documents and other collections.

LANGUAGE CLUSTER

A Language Cluster is a collection that functions as a multilingual document. If you insert two or more different texts into a cluster, each in a different language, the text that corresponds to the language the user has set will be displayed when the cluster is accessed.

SEQUENCE

A Sequence is a collection where, in addition to the collection listing, the system also generates Next and Previous links between the member elements. This means that users need not press the Back button on their browsers or select the next item manually when they want to look at all the elements one after the other. Sequences are useful in situations where a set of documents (pages) should be visited in a certain order, such as in many Web Based Training applications.

Figure 92: The Sequence structure element EXAMPLE OF A SEQUENCE

Figure 92 shows the logical structure of a sequence of 7 HTML documents. As an example of how easy it is to maintain such a structure on Hyperwave Information Server, consider the following scenario: We want to insert a new HTML document between documents 4 and 5 (see Figure 92). With a server that does not maintain links automatically (hyperlinks are statically stored within the HTML documents), we would have to perform the following operations:

Edit the table of contents document and insert a hyperlink to the new document.
Edit document number 4 and update the Next hyperlink to point to the new document instead of the (old) document number 5.
Edit the (old) document number 5 and update the Previous hyperlink to point to the new document instead of document number 4.
Edit the new document and insert hyperlinks to

a) the table of contents (Up),

b) document number 4 (Previous), and

c) (old) document number 5 (Next).

Copy the new document to its correct place on the server

In other words: in addition to physically moving the new document to the server, we would have to create or update 6 hyperlinks in 4 different documents. Needless to say, this is a lot of effort, and there are a lot of opportunities to introduce errors or inconsistencies. The same effort is required when removing a document. Also, it means that whoever is supposed to insert information into the sequence must have write access rights to all other documents of the sequence and the table of contents.

With Hyperwave Information Server, we would perform only one operation:

1. Insert document into sequence

This operation physically copies the document to the server, and inserts it into the sequence, taking into account the sort order defined for this sequence (e.g., "alphabetically by title of the document"). The rest (maintaining the link structure) is done automatically. In fact, it is not done at all on insertion, but the links are generated dynamically when users access the sequence or its elements. When you want to remove a document from a sequence, you just delete it. The navigation is adapted automatically.

Obviously, this reduces maintenance effort significantly. Additionally, the dynamic generation of hyperlinks between the elements enables a number of features not possible with static HTML:

The identical documents can be members of a number of sequences ("multiple views"). The Next, Previous, and Up links are adjusted depending on where (what sequence) the user came from. There is no need to create physical copies of documents.
Similarly, you can use the sequence within another sequence. This means that your sequence becomes a component that may be re-used, in different contexts.
Access permissions are taken into account. When a user is not allowed to see, e.g., document number 3, the Next link on document number 2 points to document number 4, and the table of contents would not list document number 3. The same would be true if document number 3 had expired. In other words, the user would never know that document number 3 even existed.
Every user who has been granted write permission to the sequence object can insert information into the sequence, without having to modify other people's documents. This means that people can collaboratively and securely work in the sequence.

Note: The features listed above do not only apply for sequences, but for any class of container.

If you insert documents into a sequence, then when a member of the sequence is accessed, buttons appear which let you navigate sequentially, forwards or backwards, through the members.

MULTICLUSTER

All documents that have been inserted into a MultiCluster are displayed as a composite document when the MultiCluster is accessed.

While both collections and sequences display a menu of the elements contained, the MultiCluster combines all its elements into one document and displays the result. As with other Hyperwave Information Server functions, this is done dynamically when a user accesses the MultiCluster, so that sort order, expiration time (see page 148), access permission, and user preferences are taken into account.

MultiClusters are useful in situations where you want to combine relatively small pieces of information (paragraphs, images, etc.) into one document. The individual pieces can be supplied by different authors (so that users can collaboratively create the large document without getting in each other's way), can have different access permissions (e.g. some paragraphs can only be seen by privileged users), and can be reused in other contexts.

EXAMPLE OF A MULTICLUSTER

Let us look at a simple example of a MultiCluster: a calendar of events:

Each event is a small document, either created by a customized entry dialog or an HTML editor. The "expiration date" attribute ( page 148) is set to the date of the event. The event listing is a MultiCluster, which is sorted by expiration date. This produces a single event listing sorted by date of the event. Listing of old events is automatically suppressed by the system (because expired objects are not displayed) without human intervention, so it looks as if the listing were maintained daily. Note that users with write access to the MultiCluster may add events, and edit or remove "their" events; there is no need to have a dedicated Webmaster to edit the event listing, or to grant all users write access to the whole listing.

ALTERNATIVECLUSTER

The AlternativeCluster is similar to the MultiCluster, but instead of returning a combination of all elements, the AlternativeCluster returns only one of its elements. An alternative cluster is a collection of documents from which exactly one document is chosen by certain criteria. The behavior of such a cluster is more like that of a simple document than a collection. The selection of this element is based on matching user preferences with attribute values of the elements.

Two typical applications for alternative clusters are

To allow a choice between documents which have different levels of quality but the same type (MimeType). A typical application are images with different resolutions or color depth which represent the same content, but consume different amounts of space. A user with low network bandwidth is able to receive an image with less quality in a reasonable amount of time.
To allow a choice between documents in different representations which do not have the same type (MimeType). Sometimes it is necessary to offer information in different data formats or media types. For instance, an announcement may be available in PostScript, HTML, sound or even video format. Users can set their preferred document type.

While it is in principle possible to select the element based on any attribute, the following types have proven to be useful:

EXAMPLES

Select by language: Use the user's language preferences to decide what document to show. In this case, the AlternativeCluster contains a number of translations of the same information, which is quite useful for managing multilingual applications.
Select by MIME Type: Use the user's preferred MIME Type settings to select a document. In this case, the AlternativeCluster contains the same information in different encodings (MIME Types), e.g., MS Word, PDF, and HTML. This is useful when the users do not have the same functionality at their client terminals (e.g., some have UNIX workstations where MS Word is not available), or need to perform different tasks with the information.
Select by Quality: Use the user's preferred Quality settings to select a document. In this case, the AlternativeCluster contains the same information in various encodings or qualities (specified by the Quality attribute). For example, an image may be available in lossless compression (e.g., GIF) and various quality levels of lossy compression (e.g., JPEG). This is useful in situations where the users have different bandwidth available (e.g., modem vs. LAN) and want to trade quality for speed.

Rules for choosing a document from an alternative cluster:

If the alternative cluster has a MimeType set then only documents of this MimeType are taken into account.
These documents are compared to the PrefMimeTypes (preferred MIME types which are stored in the user record) from left to right. Documents with the best matching MimeType are kept for further processing.
From these documents (all with the same MimeType) the ones with the smallest difference in quality compared to the quality given with the matching MimeType in the PrefMimeTypes field are selected.
If the resulting set contains more than one document, the one with the best matching language is chosen.

Alternative clusters can be nested.

COMBINATIONS

All kinds of combinations of structural elements are possible (while some combinations are more useful than others). For example, an application might be built as a collection of sequences of MultiClusters of AlternativeClusters of documents. Figure 93shows another example.

Figure 93: Structure elements combined with hyperlinks REDUCTION OF HYPERLINKS

Figure 93 also shows an important philosophical aspect of building applications with Hyperwave Information Server. The bold arrows represent "real", referential hyperlinks, which may of course be combined with structural elements. The goal is to replace the structural (organizational) hyperlinks by a combination of structural elements as much as possible, and use hyperlinks only for cross-references. This has the tremendous advantage that the system itself takes care of generation and maintenance of structural hyperlinks, with all the added benefits of dynamic link generation. It also means that the site can easily be restructured later, without editing any documents.

Benutzer: gast • Besitzer: hwsystem • Zuletzt ge�ndert am: