FRANK'S FALLACIES, FACTS, AND FICTION

Hello. As most of you know, I am writing an article each month 
for the LANET News. This activity helps to keep me off the 
streets at night and acts as a safety valve to keep me from 
uttering too many profanities in the course of dealing with 
mysterious entities like Netware, DOS, and Windows. I always 
welcome feedback of any type and will incorporate any comments, 
suggestions, and flaming responses without any mention of the 
identity of the respondent, unless given the specific written 
permission to use her or his name. I can be reached via:

1)e-mail at: fchao@cerfnet.com
2)fax at: 800-316-8206
3)voice mail at: 213-600-1235
4)messaging via the LANET BBS  (more on that later)


FICTION ??

The rumor mill keeps dangling the one about Windows 95 being 
delayed until 96. The alternate rumor (antidote) is that 
Microsoft will shed many features/modules from Chicago in order 
to get in out this year on time. I know several brave souls who 
have been using the latest beta both at work and home. One of 
these Microsoft aficionado's keeps telling me how great it is. 
If he starts screaming about having problems with the latest 
Chicago beta, I will make sure that you hear about it here.  

Last Thursday, a LANET member showed me a recent Network World 
article that states that Novell has decided to offer a 2-user 
copy of Netwarer 4.1 as part of a Software Developer's Kit (SDK) 
for less than 100 dollars. I will put on my Dick Tracy hat and 
check this one out. Keep sending me those juicy rumors--your 
anonymity is assured. 



FALLACIES ??

Don't let all that talk by the Novell folks fool you. You've 
undoubtedly heard all their claims about how Personal Netware 
is a "Universal Client":



One of my problems with that claim is that Personal Netware will 
not communicate in any way with PC that are running Netware Lite 
and acting as servers. That means that, if your unsuspecting 
client "upgrades" a Netware Lite PC to Personal Netware, all the 
PC's that that particular PC shares resources resources have to 
convert at the same. Don't make the mistake that one of clients 
did: she converted her PC from Novell Lite to Personal Netware 
and became an island unto herself--she was instantly shut 
out of access to the all the PC's on her LAN that still were 
running Netware Lite. Some truth in packaging and in 
advertising on the part of Novell would have been appreciated: 
To most of us, a "universal" client is one that operates with 
every piece of server or peer-to-peer software in at least 
Novell's lineup of products that are or can act as servers.

My second problem with Personal Netware is related to my lack 
of "customer satisfaction" with VLM's in general: Somewhere 
around the 15th of March this year, a CNE who I greatly respect, 
told me to "beware of VLM's" and I thought that his "significant 
other" had been dragging him off to too many Shakespearean 
productions. However, he is absolutely right:

Back in the pre-VLM days, the NETX DOS shell/requester of 
Netware 3.11 and the client.exe and server.exe modules of 
Netware Lite were stable as a rock. Back then and to this day, 
they make Netware PC clients a wonder to behold--transparent 
and efficient, with little in the way of excitement for the 
Netware administrator or end-user. Then, desiring to remain 
"in vogue", "current",  etc.,  I started converting PC clients 
to the VLM-based stuff: Personal Netware and the current version 
of Netware 3.12's client and all sorts of poop hits the fan: 
I have yet to find a 386 or 486 which will run VLM's with either 
external or internal cache turned on. So Novell says that they 
don't use cache. They don't have to trash it in order to make 



their point. What gives? Are they trying to outdo Microsoft by 
showing the world that they are big and strong enough to make 
grown men turn off their caches? Speak up if you have a 
patch/fix/workaround that have not stumbled on yet--I can only 
squeeze in so much plug, play, pray, and cuss time each week.  

If you are amongst the large numbers of good folk out there that 
are still using DOS applications, and you "upgrade" to VLM's 
because it is the latest thing, you will find that you can do 
less in the way of "shell" type exits to DOS:  To illustrate, 
when I run Netware Lite on a 486, I can fire up Procomm Plus, 
dial up into a bulletin board, hit alt-F4 to jump to DOS, and 
then fire up Xtree Gold. When I run Personal Netware and I try 
this trick, I end up with an "out of memory" message when I get 
the Xtree Gold part, in spite of having 8 meg of RAM, loading  
everything high, and running memmaker. 

To illustrate further, when I run Netware Lite on a 386, it 
generally acts like Netware is not even there. However, when I 
run Personal Netware with it's concomitant VLM's on a 386, with 
no additional TSR's (memory resident programs) loaded, I cannot 
even run MS-DOS edit.com on a 100k text file without incurring 
an "out of memory" error. This error effectively keeps me from 
editing many of my files with DOS edit when running VLM's on a 
386. 

I have not been able to see how upgrading to VLM's does anything 
for Windows 3.1 either. Like DOS, Windows 3.1 seems happier with 
the smaller footprint of Netware Lite. Ramwise, there ain't 
nothing virtual about VLM's.

Also, I have not met a humble 8088/XT that does not get 
indigestion from VLM's--just bring a VLM anywhere near one, and it 
will lock up to the point that even the three-fingered salute to 
the keyboard will not revive it. If you still have XT's on your 
network and have also tried to force-feed them VLM's, please drop 
me a line. 


While helping a client migrate from Netware Lite to Personal 
Netware, they actually lost functionality: In a Netware Lite 
system, when a PC acting as a server is turned off and then 



powered back up again, the clients that were logged into the 
"server" are smart enough to re-establish their connections. With 
Personal Network, the VLM's are actually dumber. When a PC acting 
as a server is powered down and rebooted, the VIM-based clients 
that were previously attached to this server are not intelligent 
enough to re-establish their connections. Instead, the user must 
type in

vlm /u 
vlm 
startnet.bat 

in order to re-establish their login connections to the 
"server". My client is not a happy camper and I felt about 2 
feet tall after explaining this loss of functionality to her. 


So what does upgrading to Personal Netware really buy us? The 
fabulous right to connect to the Netware 4.X that my boss might 
finally get around to buying in the year 2000?! In other words, 
from my point of view, it does not buy me anything--I am sure 
that it is financially beneficial to Novell, however.

As an interesting digression: the ability of the client side 
of Netware Lite 1.1 to re-establish connections to PC's acting as 
servers makes it a remarkable piece of code. I have not seen this 
capability in any other network operating system from Novell. In 
fact, I have not seen this capability in the Unix or Banyan 
environments, either. It is definitely something that Novell did 
right. This feature makes the network operating system so 
transparent that many of my clients that are running Netware
Lite forget that they are running a network since their mappped 
drives act so much like local drives. 



FACTS ??

The non-network version of Xtgold will not re-name subdirectories 
on remote drives. That is the first thing that I have found that 
it will not do. In this Windows-based point and click world, it 
is a bit hard to teach endusers, but the network and non-network 
version remain a great time-saving tools for network 
administrators.
 
After fiddling with about a dozen database systems for 
IBM-compatible PC's--both DOS and Windows versions--for the last 
5 years, I remain a firm believer in the superiority of Paradox, 
in it's various flavors. Don't get me wrong, however. I will use 
whatever database system/software that the customer wants and 
loves. But, if given a choice, Paradox still is tops in my book. 
My only complaint is that Paradox for DOS versions 3.5 and 4.0 is 
purposely LAN unaware. In order to access files on a "mapped" 
network drive, one has to fork out bucks to Borland for a LAN 
license. Foxpro, Dbase, and Microsoft Access do not have these 
limitations. Now that Novell is selling Paradox as an add-on to 
the PerfectOffice suite, I hope that either:

1) They are pushing/selling the network version of Paradox
or
2) They use their clout to get Borland to remove the silly 
   network-hostile restrictions from the non-network versions 
   of Paradox. 





Until Microsoft, and the "big" banks get their act together and 
provide standards, industry consortiums, etc., reliable check 
service/payments via computer/modem remains a "dream" for the 
future. I tried to turn the dream into reality last year and got 
bit for trying the bleeding edge. Allow me to relate my 
experience: 

"Checkfree" is a bill/check payment service that is accessed 
by modem. Off-line, using their proprietary software that is 
loaded on your IBM-compatible PC off-line, you tell it who to pay 
and how much. Then their software dials a local number which 
happens to be your local CompuServe phone number and Checkfree 
uses your bank's checking account to pay the desired payments: 
They figure out the form of payment--electronic funds transfer 
or paper checks. After trying "Checkfree" service for 4 months 
last year, I gave up in disgust. Every month, 3 or 4 of my 40 or 
so electronic "checks" would get lost in the "system". Using the 
"messaging" feature of Checkfree software, I would request that 
Checkfree investigate the "lost" payment. They would reply that 
they would start an investigation immediately. Several days would 
pass. They would they generate a reply that my only recourse was 
to wait and hopes because the "problem" was with the ultimate 
payee. Despite their promise in their marketing literature, 
Checkfree was unable and unwilling to send me a paper letter to 
"prove" that they had paid out the "lost" payment. In the 
meanwhile, any phone calls to the ultimate payee would indicate 
that they were not paid yet. Any phone calls to my bank would 
invariably yield the information that the electronic "check" had 
"cleared", as far as they were concerned. Any calls to either 
the bank or the ultimate payee would invariably yield questions 
like "What is Checkfree?" and I would end up wasting time 
educating the customer service persons on how the system worked. 
All 3 of the months that I tried to use "Checkfree", the local 
phone company posted my payments 6 to 8 weeks late, various 
credit card payments were posted between 8 and 20 days late. 
After 3 months of this nonsense, I woke up one morning with the 
uninspiring thought that the service, was costing me more time 
than it was saving me, and I promptly mailed them a letter in 
order to terminate the service.  To be fair, "Checkfree" is 
usable if you are willing to "crutch" it in one of two was:

Workaround 1: make extra payments ahead of time each month to 
utility companies, credit card banks, and creditors that have 
long delays in processing Checkfree payments so that these 
entities do not ever get to the point of dunning you for 

nonpayment--I have read in PC Week that some wealthy folks 
are willing to pay in "pad" payments to compensate for such 
problems with Checkfree; however, I am neither willing nor 
able,

Workaround 2: use regular paper checks for payees who have the 
impossibly long processing delays for Checkfree payments and use 
Checkfree for those vendors that are more cooperative.


Like workarounds in the computer world, it is up to you to 
decide whether the less-than-optimal situation is worth doing 
and for me it is certainly not worth horsefeathers so and my 
dream of doing checking by computer remains just that for now.


(Warning: The following is a bit more sarcastic than usual, 
even by my  standards ;) 
(Note the "emoticon".) 
Doing purchases on the Internet saves me money but not a way 
that you might think: In my weaker moments, I have to tried to 
buy darn near everything on the Information Superhighway. What 
happens is that about a third of all the orders that I place end 
up in cyberspace "never-never land" and never became real world 
transactions. All kinds of companies from humongous corporations 
to little roadside holes in the walls have managed to lose my 
orders. I always log/capture the entire attempted order via a 
log/capture text file and store these as a permanent record on a 
floppy, so I can prove that I am not making all this up. I have 
no idea where all these orders go. They just disappear into the 
ether. In other words, doing commerce on the "Net" remains kind 
of flaky and "iffy".

Every time I send off an article to the editor of this rag, I 
get a twinge of guilt. I think about how I am contributing to 
the killing of trees and spotted owls in some pristine forest 
somewhere. However, it also reminds me of some stark realities: 
The proportion of people with even slight computer literacy 
remains small and the proportion of people who are modem 
literate are in turn a fraction of the first fraction. Those of 
us who work with desktop computers often delude ourselves into 
thinking that everybody out there is computer literate and modem 
literate. It ain't so. I know quite a few degreed computer 
engineers who don't have a clue as to what to do with a modem.   

Think about all the people you know. What percentage are 
computer and modem literate enough to turn on an IBM-compatible 
or Macintosh "PC" and log into any BBS?  Amongst all of my 

friends, relatives, acquaintances, and creditors, the 
fraction of PC-literacy is about five percent and only about a 
quarter of that fraction are also modem-literate. In other 
words, only about 1 percent of the people that I know know how 
to use modems. 

Part of the reason for this is the fact that no one ever 
really improved on the archaic Hayes command set. It was 
invented about 15 years ago by Dennis Hayes of Atlanta, Georgia 
and by today's standards, in a point and click world, it is about 
as user-hostile as anything out there. Sure, the Datastorm and 
DCA folks have written GUI-based software such as Procomm for 
Windows and Crosstalk for Windows that hide most of the
complexity from the unsuspecting end-user, but when the chips 
are down: during installation for such software packages, when 
things are acting flaky, she who does not know some of the basics 
of the Hayes command set will be stuck.

And don't even suggest that modems and modem software is now so 
darn reliable that the end-user does not have to concern herself 
with troubleshooting. The people who pay me to have all this fun 
have given me the opportunity to experiment with virtually every 
modem software package and over 2 dozen brands/models of modems 
over the past 4 years and I never ran into a single modem or 
software package that not act up fairly often. If the access to 
the Information Superhighway for 99 percent of us is a dial-up 
modem via a desktop computer, then for 99 percent of us Internet 
peasants, the metaphorical on-ramp is a bumpy dirt road strewn 
with boulders.  

Because modems are the chief way of access onto the Internet, 
these same minuscule percentages translate into an extremely 
small percentage of persons who actually access the Internet. 
Even if one includes the "Big Boards" like Compuserve, 
Prodigy, and America Online into the equation, from the 
end-user perspective, the "Internet explosion" that the press 
keeps hyping about might consist an increase in percent of the 
population accessing the Internet from 1 percent to 2 percent. 
In otherwords, the actual percentage of persons accessing the 
Internet is miniscule--the "explosion" if more of a media hype 
sort of thing than a mass migration of the populace. 



Also, based on my observations, of this small group that 
actually access the Information Superhighway, my guess if that 
more than 2/3's of these folks access it at work and cannot hit it 
from their homes. 

It is often hard to believe these realities, but if you are 
reading this rag, you are probably amongst the small percentage 
of people who are surrounded by and thus have easy access to 
desktop computers of some sort. 

The lack of competence in modems may be at least a partial 
reason for why our LANET BBS continues to have very little in 
the way of utilization. I logged in one whole time during the 
month of February and it appears that I was the only LANET 
member to do so. For the second month, I think that I am the 
only LANET member to log into the BBS. Apparently, that was 1 
more time that the sysop 
did. I am beginning to suspect that our BBS's phone number is too 
darn expensive for most of our membership. Don't know the
solution. The Orange County chapter (OCNET), of which I am also 
a member, has a thriving BBS called "Cyborg". Perhaps they could 
bottle whatever they have and sell us some. 

Finally, I finally got off my butt and sent a check to the 
address that our editor gave me, in an attempt to renew my 
membership in LANET. I have not heard from anyone yet, so I am 
probably still an "undocumented" member. In comparison, when I 
sent in my renewal for the Orange county chapter, they sent me a 
pile of literature and a nice letter acknowledging my payment. 
We could also probably use some of their marketing smarts. On 
the positive side, our meetings are as informative as anyone's--  
each meeting is a learning experience that I cannot get while 
doing all that "routine" "fire fighting" in the course of 
generating my income stream.

Besides, after my move back to an area of L.A. county where salt 
spray from the ocean corrodes cars, I cannot make their meetings 
without chartering a private helicopter.





Many thanks to those who have provided me with tidbits and 
comments. Hope to see you next month.


(The previous diatribe is solely the private opinion of the 
author, and neither LANET, SCNUI, NUI, Novell, the author's 
unsuspecting employers and clients, nor anyone else for that 
matter vouch for the validity for what was stated.)