Adobe(TM) Acrobat(TM) Search for UNIX Read Me Notes for Version 2.1

This Read Me file contains last-minute product information for Acrobat 
Search 2.1. The Acrobat Search online guide (Help-Search.pdf) will also provide 
essential information to help you begin using Acrobat Search 2.1.


Known Issues with Search
========================

On Solaris(R) 2.3 or 2.4 or SunOS(TM)
-------------------------------------

A problem exists in Acrobat Search on these platforms that prevents it from 
connecting to an index where the index directory and subdirectories are in
all uppercase letters. This means that you will be unable to connect to any
index copied from an ISO9660 CD-ROM in a method that maintains the uppercase
characteristics of the ISO format (e.g., the index is copied from an HP-UX or 
Macintosh machine to a server). 

There are 2 ways to solve this problem:
(1) Force lowercase the index directory and its contents. For example, if the
index is named FOO.PDX, then the index directory is named FOO. The index file
name and the index directory name and its contents (which is multiple 
directories and files) all need to be changed to lower case.

(2) Make symbolic links which are lowercase to the index and index directory and
all its contents.

When reading directly from a CD-ROM on Solaris 2.4, do not mount the CD-ROM 
with the -nolowercase option.

On Solaris 2.5
--------------

As of the ship date of Acrobat 2.1 for UNIX, Acrobat Search does not run on
Solaris 2.5 (Beta). Acrobat Exchange or Acrobat Reader
with Search will not boot on Solaris 2.5. To get around this problem, you must
remove the Search components from the <install dir>/<platform dir>/plug_ins 
directory. Those files are: acrosrch.uid and acrosrch.api.

Using Indexes on a Novell Server
--------------------------------

Due to a Novell permissions issue, it is not possible to search an index which
is being built by Acrobat Catalog for Windows or Macintosh. It is necessary
to wait until the index is finished being built or the following error message 
will appear: "No documents were found that matched your query."

Miscellaneous
-------------

1. If two indexes exist in the same directory, one in all-uppercase and the 
   other in all-lowercase naming, only the lowercase index is recognized and 
   searched (even though both indexes can be attached). Do not have 2 indexes of
   the same name in the same directory.

2. Wild Card searches with the "Match Case" option checked will fail if the 
   index was not built with the "Case Sensitive" option. The user will receive
   the message: "No document found...". Uncheck "Match Case" and try the search
   again.

3. If an index (*.PDX file) is renamed or removed after being added to the 
   Index list, it is still available for searching. Searches will proceed 
   correctly as long as the directory corresponding to the original *.PDX file 
   has not changed. However, the info about the index will be incorrect. We
   recommend not moving *.PDX files or changing their names if searches might
   be in progress.


Opening Files in the Search Results List
========================================

Slow to open or file not found
------------------------------

When opening a file from the results list it may take a long time to open the 
file or in some cases the file can not be opened at all. This problem is caused 
by using non 8.3 filenames and directory names with more than 8 characters. To 
resolve this problem, name your files and directories using 8.3 conventions and 
rebuild the index. You can also avoid this problem by checking "add ID's to 1.0 
files" in the index options dialog, and using 8 character directory names for 
your top level directories. These changes will help search locate and open files
more quickly.

When searching for common words, such as searching for "the" by itself, search 
may appear to hang. In fact, Search is just taking a very long time. This 
happens because the number of hits is so large it takes a while to retrieve them
from the index.

File has been changed dialog when the file has not changed
----------------------------------------------------------

Novell date reporting can be inconsistent between platforms. This may cause 
Search to erroneously report that a file has been changed since the index was 
last built. This happens when the indexing machine puts a file modification
date into the index that differs from what your system sees in the file system. 
Search will complain that the file has changed when in fact it has not.


Supported Networks
==================

Acrobat Search and Catalog may have problems indexing or finding documents due 
to differences in how networking software maps non 8.3 names. For example, what
is seen on the network by a UNIX machine may be different from what is seen by a
PC. The following networks have been used in testing Acrobat Search and Catalog:
 - Novell Netware & Netware Lite 
 - Banyan Vines 
 - Sun Select PC-NFS 3.5, 4.0, 5.0 and 5.1 
 - NetManage ChameleonNFS 
 - IBM OS/2 2.1 
 - Microsoft Windows for Workgroups 3.11 
 - Microsoft Windows 95 
 - Microsoft Windows NT Advanced Server 3.1

Other networks may work, but have not been thoroughly tested.


Unexpected Results
==================

No results found
----------------

If you are searching for a phrase which contains a stop word you may 
unexpectedly get no results found for your search. For example searching for 
"nick of time" will lead to no results found if "of" is a stop word.

Files not appearing in search results list

If proximity is turned on, files which contain the words being searched for may 
not appear in the results list if these words are 2 to 3 pages apart. For 
example when searching for "income and report" if both words appear in a file 
with 8.5 x 11 pages full of 12point text then the file will only appear in the 
results list if income and report appear are on the same page or consecutive 
pages. Some words will not be found if they contain high ANSI characters such
as ligatures and are in Type3 fonts. For example you can not search for 
"AEsop's Fables" if they are written in a Type3 font because it contains the 
high ANSI character AE.

Incorrectly finding no results
------------------------------

If the index has been purged and rebuilt since your last search you will need to
detach and reattach to that the index. Searching the index without reattaching 
will not find any documents even if the index contains hits.

Unexpected words being highlighted
----------------------------------

Using some of the word options such as sounds like and thesaurus can lead to 
searching and highlighting of unexpected words. For example the sounds like 
feature is designed to work best with proper names. For other words sounds like 
will not always give the results one might expect. Use the words assist dialog 
to understand what words are being searched for when you use word options.

Search terms only being highlighted once per document
-----------------------------------------------------

When using the proximity word option only one occurance of the search terms will
be highlighted in each document even if the search words occur multiple times. 
The proximity option sorts the results list bases how close together the search 
terms are in a document. The document is ranked by the search terms which appear
closest together. Only this occurance of the search terms is highlighted in the 
document. The other occurances are not highlighted because they did not 
contribute to the document's high ranking in the results list.

Proximity not being used
------------------------

Proximity is only available for simple AND searches such as "earnings AND 
report". Complex searches such as "earnings AND (report OR analysis)" will 
automatically turn off proximity.

Indexes
=======

Indexes unavailable on UNIX file servers
----------------------------------------

See above in "Known Issues with Search on Solaris or SunOS".

Incorrect dates in index info dialog
------------------------------------

The creation and modification dates of indexes shown in the index info dialog 
may appear incorrect. This will happen on a Novell server or an index created 
on a PC or Macintosh(R) and copied to a Novell server.


Miscellaneous
=============

Network server overload
-----------------------

A network file server can reach max utilization when many users are running 
broad searches at the same time. This causes an overload on the server slowing 
down all other processes. This occurs when many broad searches attempt to read 
the entire index which forces the server to load the entire index into its 
cache. As indexes can be very large the caches can be overloaded. To avoid this 
problem discourage broad searches such as "*report" which will read the entire 
index.

Indexes not supporting word options
-----------------------------------

Some indexes are built without the support for word options such as stemming or 
thesaurus. The documentation incorrectly states that when searching against 
these indexes with these options Search will not find stems or synonyms of the 
search words. Search will in fact find these even when these options are not 
included in the index. Not including these options in the index only effects 
the word assistant which will not be able to show stems or synonyms for this 
index.

Internal Error while searching
------------------------------

When searching documents which contain long titles you may get an internal error
message from Acrobat. This message is due to having titles which are same 
through the first 63 characters in the title. For example, suppose we have two 
documents with the titles: "The Quick Brown Fox Jumped Over The Lazy Dogs. By 
John A. Smithson" and "The Quick Brown Fox Jumped Over The Lazy Dogs. By John A.
Smithsun". These titles are 66 characters long and are identical in the first 63
characters. The dialog will appear in this instance if the search results are 
sorted by title. As a work-around, change all document fields (especially the 
Title) to be unique within the first 63 characters if their field values are 
more than 64 characters long and reindex the documents.


Adobe, the Adobe logo, Acrobat, the Acrobat logo, and PostScript are trademarks 
of Adobe Systems Incorporated or its subsidiaries and may be registered in 
certain jurisdictions. Macintosh is a registered trademark of 
Apple Computer, Inc. HP is a registered trademark and HP-UX is a trademark of
Hewlett-Packard Company. Motif is a trademark of Open Software Foundation, Inc. 
Solaris is a registered trademark of Sun Microsystems, Inc., which has not 
tested or approved this product. Sun and OpenWindows are trademarks of Sun 
Microsystems, Inc. SPARC is a registered trademark of SPARC International, Inc. 
SPARCstation is a registered trademark of SPARC International, Inc., licensed 
exclusively to Sun Microsystems, Inc. and is based upon an architecture 
developed by Sun Microsystems, Inc. UNIX is a registered trademark in the 
United States and other countries, licensed exclusively through X/Open Company, 
Ltd. X Window System is a trademark of the Massachusetts Institute of 
Technology. All other products or name brands are trademarks of their 
respective holders.

copyright 1983 - 1995 Adobe Systems Incorporated. All rights reserved.