. 33
( 87 .)


Figure 10-1 URL mapping administration.

contexts. To add a new context, click New Context and provide a label such
as ¬nance. If you wanted to add some more child contexts, you could then
click Finance and create another context. However for simplicity, we are
leaving it at one level. The label is appended to the portal pre¬x, which
in this example is http://sandbox1.rigorconsultants.com/wps/
myportal/. After you create the context, click the icon Edit Mapping and
select the portal page that you want to be associated with the label and click
OK. If there are too many pages, use the search capabilities. Choose the ¬eld
you want to search (such as title) and then enter a value that will ¬nd the
page you want to choose. If you need to change the label, click the icon Edit

Custom Unique Names
Custom Unique Names allows you to associate a human readable name
with the object IDs created by the portal. These object IDs are cryptic so that
portal resources can be identi¬ed uniquely between different portals. Portal
object IDs require human readable names since they are referenced when
exporting or importing a portal con¬guration, linking to another portal
resource, accessed by third-party vendor products, and so on.
WY009-10 WY009-BenNatan-v1.cls May 14, 2004 0:16

182 Chapter 10

To create or modify a custom unique name, click Custom Unique Names
under Portal Settings. Select the resource type that you want to associate a
custom unique name with. You can choose from the following:
Portlet Applications
URL Mapping Contexts
User Groups
Upon choosing the resource, a page will display the resource title, its
unique identi¬er, and custom name. A search bar will also be displayed
that will allow you to ¬nd the resource by title, description, keyword, last
modi¬ed, unique name, or all available. If you click the Edit icon, you can
enter a custom name for the unique identi¬er associated with the resource.

Supported Markups
Supported Markups is used to administer the markups associated with the
portal. Upon clicking the Supported Markup portlet, you will be provided
the options to add, edit, activate, show info, or delete a markup. When
you add a markup, you need to give it a name, a MIME type (the Internet
standard on how messages must be formatted), and default character set.
More information on markups is provided in Chapter 24.

Supported Clients
This section is important if your portal supports unique devices. Supported
Clients is where you de¬ne the types of devices that your portal supports. By
default, most browsers and popular wireless devices are already supported.
However, sometimes you want to add a unique device. To do so, you need to
add support for the Sony Ericsson T310 device, which is done by performing
the following steps:

1. Click Supported Clicks, which will display the Manage Clients page.
2. Click Add, and the Manage Clients page is displayed, with the
Clients ¬eld.
3. In the User Agent ¬eld enter the name of the client surrounded with
a period plus asterisk combination. This must match as close as
possible to the user agent string the client sends in its request header.
For T310, it is .*SonyEricssonT30/R201.*
4. Choose the markup from the drop-down box. These markups were
de¬ned in Supported Markups. For T310, it is wml.
WY009-10 WY009-BenNatan-v1.cls May 14, 2004 0:16

Portal Administration 183

Figure 10-2 Adding a Sony Ericsson T30/R201 client.

5. Enter the markup version (optional). For this situation, leave this
6. Enter the model or name of the client (optional). For T310 it is Sony
7. Enter the version associated with the model (optional).
8. Enter the capabilities of the client such as Frames or JavaScript. Click
Add or Delete. Don™t add anything since this is a dumb mobile
9. Choose the order from the drop-down box where the client will be
entered in the client registry. The most speci¬c user agent patterns
should be placed close to the top. Add it before R201. Your screen
should now look like the one shown in Figure 10-2.

Search Administration
WebSphere Portal provides a very sophisticated search and document col-
lection capability. Basically you can create collections of various types of
documents for other portal users to access. You create the collection using
a very sophisticated search engine that supports a wide range of Internet
WY009-10 WY009-BenNatan-v1.cls May 14, 2004 0:16

184 Chapter 10

search operators. Each collection is a set of documents that can be sum-
marized and/or categorized. Collections can be automatically categorized
using a prede¬ned static taxonomy or can be user de¬ned. Documents can
be anything from HTML to Excel spreadsheets and then can be stored on
local or remote sites. The collections can be scheduled and ¬lters can be ap-
plied to exclude or include any documents. In this section, we will discuss
how WebSphere Portal search and document collection, including admin-
istration, at a high level. Chapter 12 will discuss this in greater detail.
Under Manage Documents Collection, you can de¬ne Web sites or doc-
uments, which will be prefetched by WebSphere Portal Web crawler. Web-
Sphere Portal will create indexes that will map key words and terms to
the documents. These indexes are referenced by setting the IndexName
parameter in the Document Search portlet.
In learning how to use this feature, create a document collection from
information stored at http://www.rigorconsultants.com. Refer to
Figure 10-3.

1. Click Administration ➪ Portal Settings ➪ Search Administration.
2. Under Document Collections, click Create Collection.

Figure 10-3 Creating a collection.
WY009-10 WY009-BenNatan-v1.cls May 14, 2004 0:16

Portal Administration 185

3. Specify directory path where you want the collection to be created. If
you give it a name WP saves it as an appropriate directory under
<was_root>. Call this collection Richard.
4. Specify Collection language; select English.
5. Select Categorizor. You have a choice of none, prede¬ned, or
user-de¬ned. If you choose prede¬ned, WP will automatically
categorize your documents in over 2,300 subjects. Normally you will
choose prede¬ned; however, to show all the features you will choose
user-de¬ned. Note that when you choose user-de¬ned, the Category
Tree icon will appear on the Manage Documents Collection screen. It
will not be there if you choose prede¬ned.
6. Select Summarization. It can be either none or automatic. If you select
automatic, WP will create a coherent narrative summary for each
document, independent of language.
7. Check the Remove Common Words from queries box and then click

Now, you have to create a custom categorization. Usually you set these
up if you want to see all the documents associated with your product. In
this case, you are going to create a category (surprise!) called WebSphere,
which will ¬nd all Web pages that contain the word WebSphere. Refer to
Figure 10-4 and perform the following steps:

1. Under Manage Document Collections, click Category Tree.
2. Under the root category, input WebSphere Portal as the sub-category
name and click Create.
3. Expand the root tree on your left and highlight WebSphere Portal.
4. Under the Manage Category Rules Box in WebSphere Portal, click
5. A new page will appear. In the Create Category Rule box, perform
the following tasks:
a. Input WebSphere as your rule name.
b. Click Apply Rule to Content.
c. Input WebSphere as the key word that you want to select
documents with.
d. Click Create.
6. Click the WebSphere Rule now appearing in the Manage Category
Rules box.
7. Click Manage Rules.
WY009-10 WY009-BenNatan-v1.cls May 14, 2004 0:16

186 Chapter 10

Figure 10-4 Creating a category tree.

Next, as seen in Figure 10-5, you need to specify from where documents
are to be collected by following the steps given below:

1. Under Manage Documents Collection, highlight the collection.
2. Under Sites in Collection, click Create Site.
3. Enter the URL of the site. For this example, it is 5. Specify the level
of linked documents to collect. This is the maximum number of
levels the crawler will follow.
4. Choose number of linked documents to collect. This is the
maximum number of documents that will be indexed.
5. Select the time to stop collection. The value of this ¬eld determines
the maximum time the crawler will run in a single session.
6. Choose Stop Fetching a Document After (sec). This ¬eld de¬nes the
maximum time limit in seconds for completing the initial phase of
the HTTP connection so that the crawler does not get stuck
in¬nitely in a bad connection.
7. Select Link Expire After (days). Each document is time stamped.
This ¬eld determines when documents will be removed from the
WY009-10 WY009-BenNatan-v1.cls May 14, 2004 0:16

Portal Administration 187

Figure 10-5 Manage Document collections.

8. Choose Remove Broken Links After (days). This ¬eld tells the
crawler how many days to wait before removing documents that it
found before but can no longer ¬nd because of invalid links.
9. Choose the Number of parallel processes. This is the number of
threads that the crawler uses during a session.
10. Enter the default character encoding. If the crawler cannot
determine the character set of a document, it uses this character
11. Click Always Use Default Character Encoding if you want the
default character set at all times.
12. Click Add All Documents to Collection Automatically if you want
the documents put in their destination folders and indexed,
otherwise they are put in the pending documents folder until an
administrator approves them.
13. Click off Robot.txt if you do not want the crawler to observe the
restrictions speci¬ed in the robot.txt ¬le.
14. Enter the proxy server and socks server if the crawlers uses on.
15. Click the Create button.
WY009-10 WY009-BenNatan-v1.cls May 14, 2004 0:16

188 Chapter 10

For this particular example, you used all the default values except In-
crease the Stop Fetching a Document to 20 seconds and you clicked Add all
Documents to Collection Automatically. If you did not click Add all Doc-
uments to Collection Automatically, then the documents fetched would be
held until an administrator released them by clicking Pending Documents
(under Manage Document Collection). We also told it to obey Robot.txt,
which is a ¬le on Web sites that informs Web crawlers if a page should be
excluded from the collection.


. 33
( 87 .)