A cynic's review of WAIS

David Steffen steffen at mbir.bcm.tmc.edu
Wed Oct 2 15:25:49 EST 1991


              A Cynic's Review of WAIS

************************************************************
*                 EXECUTIVE SUMMARY                        *
*                                                          *
*   As Rob Harper has said, WAIS is a nifty way of         *
* searching remote databases, and may well represent the   *
* wave of the future.  However, my experiences with        *
* WAIStation 0.6.2 for the Mac suggest that this           *
* program/system/whatever may not yet be ready for the     *
* working biologist.                                       *
*   If you have a remote database that you positively must *
* search, which is accessible in no other way, or if you   *
* just like to play with new toys, go ahead and jump in    *
* now; but be prepared for some work and frustration.      *
* Otherwise, leave it to the computer pioneers to get      *
* things going before jumping aboard.                      *
************************************************************

Disclaimer:  All this is only my opinion; Baylor has nothing 
to do with any of it.  I am not an expert on WAIS; consider 
this a Pournellesque review.  If you insist on the liberal 
use of IMHO, hit "N" now.  If this review makes you mad 
enough to flame, flame yourself.  If you  flame me and as a 
result I do something rash, my wife and kids will sue you.

======================== INTRODUCTION =======================

  I have to confess: computers are a serious addiction of 
mine.  Further, I have felt for a long time that something 
other than Usenet newsgroups was necessary for taking full 
advantage in the promise implicit in the vast resources 
potentially available on Internet; something that saw data 
not as transient messages, but rather as relatively stable 
bodies of knowledge.  So when one of my net.heros, Rob 
Harper, began raving about the greatest thing in searching 
remote databases, something called WAIS, and further, when 
Bionet made the archives of the bionet newsgroups WAIS-
accessible, how could I avoid jumping on the bandwagon?  My 
addiction notwithstanding, however, I am a working biologist, 
and thus I had to get some positive results with a reasonable 
effort if I was going to be able to use this new thing.  
Downloading the Mac version of the program from Thinking 
Machines Corporation (which they provide *ABSOLUTELY FREE*, 
Thanks, Thinking Machines!), guided by the clear instructions 
given by Rob Harper, went relatively straightforwardly.  
However, initial attempts to USE the program to search for 
useful database were aggravating, time consuming and largely 
unsuccessful.  Further effort has indicated to me that at the 
present time, use of the program is too complex and too 
unreliable, and resources accessible with WAIS are too sparse 
to allow me to recommend this program as a working tool.  
However, I hope WAIS really is the wave of the future, for if 
WAIS-compatible resources become more generally available, 
and if some of the bugs get worked out of the system, this 
will become a must-use technology.

======================= WHAT IS WAIS? =======================

  I am not going to try to reiterate the thorough 
descriptions of WAIS posted to the net by Rob Harper, or the 
beautiful manual provided for WAIStation *WITHOUT COST* from 
Thinking Machines, or the rather hyperbolic but informative 
article in the May 1991 issue of Byte.  However, I will note 
that after reading the information provided by all three of 
these sources, I was still confused.  This is a problem I 
frequently have when reading computer documentation; that it 
is is completely clear if I already understand what is being 
described, but totally opaque if I don't.  Nothing for it, 
then, but to download WAIStation for the Mac from Thinking 
Machine's computer, and see if I could figure the sucker out 
by playing with it.  After several hours of work I have been 
partially successful, and the following description of WAIS 
is based on that experience.

1) What does WAIS do?

  WAIStation for the MAC, *ONCE YOU SET IT UP*, logs onto 
remote databases, and allows you to search them.

2) What do you mean, "remote databases"?

  Potentially, there are lots of answers to this question, 
but I will start by talking about just one remote database as 
an example.  If you are even reading this message, you are 
probably familiar with the Bionet newsgroups.  One database, 
located on the genbank computers in California, consists of 
all of the messages ever posted to these newsgroups.

  For me, one of the most useful of the bionet newsgroups is 
bionet.molbio.methds-reagnts.  In this newsgroup, people 
discuss things like difficulties ligating PCR fragments and 
the solutions thereto, experience with different brands of 
PCR machines, different stains for detecting DNA on gels, and 
so forth.  Problem is, you may not be interested in a 
particular topic when it comes up, but 6 months later be 
desperately interested.  For example, you decide to buy a PCR 
machine and  wonder if anyone has discussed their experiences 
with different brands of machines on the newsgroup.  Perhaps 
you think you remember reading something like this, but 
didn't save the messages, or perhaps you have just 
subscribed.  All the past messages are saved on the Bionet 
computers.  This is one database.  WAIS is one way in which 
you can search through this database.  //Need an example 
search//.

3) What other databases are there?

  Several, but not so many at present useful to biologists.  
Perhaps the first I should mention is the "database of 
databases" maintained at Thinking Machines.  Thinking 
Machines maintain several other example databases on their 
machines.  However, the ones potentially useful to biologists 
are incomplete; they are included as examples only.  If you 
use Rob Harper's BioBits, all of the back issues of this are 
available on his computer in Finland.  In addition to the 
bulletin board database, discussed above, there is also a 
"biojournals" database on the Bionet computer, which I 
haven't played with yet; I would assume that access to 
Medline (which most of us have) might make the biojournals 
database less essential.  For me, the only truly useful 
databases available RIGHT NOW are the Bionet bulletin board 
database and the BioBits databases.  (Many of the other 
databases are fun, however.)

4) How is WAIS different from what exists now?

  WAIS is different from what exists now in several ways.

First: It is my understanding that it is much easier to make 
a WAIS-searchable database than it is to make other kinds of 
remote searchable databases (e.g. Medline).  In addition, I 
understand that the WAIS protocols have been placed into the 
public domain , or at least are available to those who want 
to use them.  Thus, many people will have a sufficient level 
of expertise (Rob Harper, the folks at Bionet, NOT David 
Steffen) to produce a WAIS-compatible database with 
relatively modest effort.  What this will (hopefully) mean in 
the future is that more databases will be available.

Second: WAIStation makes it easy to search databases in very 
useful ways.  For example, *ONCE YOU SET IT UP*, it is easy 
to search multiple databases with the same question in one 
step.  If you get a list of articles, only one of which is 
what you were looking for, WAIStation allows you to ask for 
"more articles just like this one" (but see below!).

============= SOUNDS GREAT!  WHAT'S THE CATCH? ==============

1) Learning how to set up WAIS can be a task!

  I get the impression that the authors of WAIStation expect 
that their software will be set up by an "expert" before the 
end user messes with it; in particular, they don't expect the 
end user to be adding sources.  (Do you have someone who will 
set WAIS up for you if you ask them to?  I don't.)  Quoting 
from the manual (the "WAIStation User Guide 0.57"):

  "Sources are generally available to everyone at a given
   site.  Questions, on the other hand, generally belong to
   individuals."

  "Your System Administrator will have arranged to have some
   or all of your site's sources updated at regular
   intervals."

In support of my suspicion, there is no information about 
adding new sources in the manual.  Thus, the first couple of 
hours I spent playing with WAIS was spent trying to figure 
out how to add bionet as a source.  After studying Rob's 
posts I thought I got it right.  No luck.  I kept getting 
error messages like:

"Connection Failed - Unable to Access Remote Database"
           -or-
"Configuration Error - Unable to Access Remote Database"

Turns out, my difficulties were aggravated by:

2) Flakey network connections!

  Decided to spend a Saturday afternoon messing with WAIS 
before trashing it.  After messing with things, SUCCESS!  
Went home.  Came in Monday and:

"Connection Failed - Unable to Access Remote Database"
"Configuration Error - Unable to Access Remote Database"

Tuesday, it works again.  It is very difficult to figure out 
the right and wrong way of doing things if the systems 
imposes errors on top of the ones you are making.

3) Nifty features give weird results!

  One of the major claims to fame of WAIS is that searching 
with keywords is just the start; once you find a reference 
you like, WAIS will find more like it.  
Use "article like/text like" with bboards gives unpredictable 
results.  Every time I do this, the list of references I gets 
gets worse, not better.  In particular, one post in 
particular, (a post to bionet.general from 
wchutt at monsanto.com on 24 Dec 90 12:16:33 GMT) always shows 
up as the best match.  Turns out this is a garbled message 
which is a concatenation of a bunch of different messages to 
different bulletin boards (many not related to bionet).  This 
concatenation seems to make it a trap for these kinds of 
searches.

  In several hours playing with WAIStation and the bionet 
bulletin boards, I have found that only the simple keyword 
searches give good results.

  Finally, it appears that not all of the advertised features 
of WAIStation work yet and that some important features were 
not included.  I can't be sure of this, as I am still 
learning to use the program.  This isn't a complaint, as this 
is still a pre-release version, but I just wanted to let 
y'all know: this is still a pre-release version.

================ I'M GOING FOR IT; ANY TIPS? ===============

Everything I say here assumes you are using a Mac connected 
to Internet via MacTCP.  If this is not you, sorry, I can't 
help.

1) When you ftp to think.com to retrieve WAIStation, you will 
find two candidates:
WAIStation-0-61.sit.hqx
WAIStation-0-62.sit.hqx
Presumably, the -62 is the more recent version of the 
program.  However, there is stuff in the -61 package which is 
not in the -62 package; you might want to download both.

Also, you might want to get wais-overview-docs.sit.hqx, which 
contains additional documentation.

If wading through all this seems like a chore and 
professional suicide, remember; I WARNED YOU!

2) The two sources you will want right away are the bionet 
bulletin boards, and BioBits.  The information you need to 
install these are as follows:

bionet bulletin boards:
----------------------

CONTACT:   Mac TCP ...
           Address: 134.172.3.160
              Port: 210

DATABASE:  biosci

biobits:
----



More information about the Bio-soft mailing list