Art Quarter Home Page

Art Quarter Fine Art Logo

Tips and Advice on Computing and Internet

Signature Dr. Werner Stürenburg, Germany

Great sites you want to check out
How to start biz on the net?
What about backup? It took me 50 minutes to save 2 GB!
Why does restore from tape take that long?
Should I save my internet downloads to tape?
How to deliver content.
How to use email links.
Experiences with dictation systems.

Q. Great sites you want to check out

Here are some places I profited from greatly:

NCSA Beginner's Guide to HTML

Primer on html

Grantastic Designs

Design tips

Profit Clinic

Small biz counseling

The Source

A Web Master's Resource Site

Associate Programs

Making Money on the Web

Word and Excel Tutorial

Computer Based Training

Kessler & Associates

Online Marketing Services

Promotion World


MarkeThink Digest

Marketer Discussion List

MarkeThink Digest

If you need more info, email me.

top | Art Quarter | feedback

Q. How to start biz on the net?

I wrote a long report about my experiences for you to learn from. There is plenty advice on design, too. If you want to read a more thorough introduction to design and promotion, see Grantastic Designs.

The bottomline is:

If you know you want to do serious business, you will need the following:

  • flat rate dial-up service
  • professional host
  • reliable mail service
  • sound page design

and some more you can learn from Jim Daniels and Corey Rudl.

If you need a great product to sell, I recommend Virtualisys. You can sign up at Virtualis.

top | Art Quarter | feedback

Q. What about backup? It took me 50 minutes to save 2 GB!

I did Novell networks since 1987 and we always had a central backup solution based on tapes. Time is no issue, you could implement a simple night start routine (server running always anyway). Of course, somebody had to look at any error messages. Then just change tape.

This is an important topic in any serious backup routine: Scheduling. We advised our clients to use 9 tapes. To not confuse things, they were labeled Mo through Fri and Week 1 to 4. So we could reconstruct any day of the week and 4 weeks back.

We always did complete backups. Backups were needed seldom, so we did not record a content database (once they were invented due to huge data) to speed up things, they piled up huge data and would have been used never.

Backups were done to assure good sleep. So the weekly tapes were distributed among employees to be kept at their homes in case of fire or theft in the main office. And to be sure, these things did occur. Our backup strategy never left any customer experience severe loss. Of course, ditributing things is some more hassle. If you need the tape you have to take a drive home. If you forget the tape needed this week, you have to drive, too.

You have to trade off convenience and security. You probably do it with all your programs right now, if they support it. If you have an automatic save option, you will set it to some sensible time, say 5 minutes. If anything goes wrong you will have lost at most 5 minutes' work.

The same with backup. The more often you backup the more hassle you will have in case of an emergency. Our clients could backup during lunch break, too, to have a copy of morning's work, so if loss occurred, it would have been at most half a day's. You get the point.

We did not think introducing a quarterly or yearly session would make too much sense. Data thus to be recovered would be too old anyway. But this may depend on the type of biz. I think tape backup is superior to all other ways in that it enables you to have these kind of backup "generations". All professional backup systems work this way. If you get yourself a DAT tape, material is incredibly cheap, so you can afford those 9 generations easily.

Actually we experienced severe problems with backup software. Old systems were comfortable (on that DOS level), they worked fine, people understood easily. Then they changed that. Newer software lacked features which are essential.

You must have an error report if you make a backup, else it's just worthless, the only thing you know is that this thing did something. I don't know the situation now. Many software famous then vanished now. Lately we used Backup Exec, too, but weren't too happy with it. But things ought to improve.

It is tempting to use the tape as a drive. In fact, this is an old hat, too. For some kind of biz like programming this could be the hit, as you had plenty of storage and could backup every few minutes. But it did not turn out that way. If you want to do that, get yourself a jaz drive, works fine, is inexpensive (tape is much less than drives, though), with compression you get 2 gigs each.

Also it is fine to transfer large data between systems. I gave hundreds of MB to a printer with a Mac system, for example. No problem (except for compression - he could not read that, had to uncompress first).

Get yourself some more drives, if you need to. I bought 3 of them for some years now, that's enough for me. I used to use it as an additional drive for a long time. Only thing annoying was the sleep mode. Would have liked to turn it off to not having to wait it speed up, couldn't be done. Now I use it seldom, so I keep it popped out as Explorer and some other software use to look there for something not comprehensible to me, hence it stops the whole system during the time it needs to speed up.

I would not recommend jaz as a backup medium, though. Backup has to be automated or else it is not done at all. My experience. You can replace any old machine, get even a better one, it's only money. But if you have to recreate your work, it's your lifetime.

top | Art Quarter | feedback

Q. Why does restore from tape take that long?

A restore time of 20 minutes for 200+ files of 50 MB is completely normal for a tape drive. That's the reason we have different technologies with pros and cons each.

A disk is a random access medium, like a music record or CD. You can pick up the needle and place it at any radius, you're at most one turn off the spot you look for. A tape is not a random access medium. I takes a lot time to just fast forward from beginning to end without even reading. If the data is at the end and your tape is at the beginning, it has to do some spinning just to get there.

If you restore lots of small data in a sequence where each is at the end and at the beginning in turn, you can estimate that the tape is spinning 99% and doing actual restoring just 1% of the time.

Actually you have the same problem with disks. If you work with them, writing and deleting, the disk will get cluttered, or fragmented. So if you write something new, it will begin at a free sector until it runs out of space, then it looks for another sector to write on.

Now imagine you habe a large image file chopped into sectors placed at the center and the periphery in turn, then the disk is busy repositioning a thousand times. So access to data is not only dependent on read time but also on repositioning time, time needed to place the head from one place to another. If your data gets fragmented, you'll spend your time on repositioning.

Cure is, of course, defragmentation. If there is a visual feedback with your defrag program, you can see very nicely how it collects fragmented data, looks for a segment big enough to hold it in one place, then writes it there. After it is done, you have no unused space between your data and a huge block of free space to write anything to. Of course, now the cycle just starts all over. If you work you will fragment your disk.

To complicate things further: disks are cut into sectors, but they only hold 512 byte (depends on the operating system). This is obviously not a good unit, as you would have to jump to thousands of sectors to read some useful amount of data. That's why lots of sectors get grouped to clusters which are taken as unit.

The number of sectors taken are again dependent on the operating system. One of the problems with the old Win systems is, for example, that there is a max number of clusters. So when disks became really big, the clusters had to grow big, too. This means that the smallest possible unit to store something into was getting larger all the time. Take an average email of 4 KB. That would have to be stored into a cluster of 32 KB. You waste 28 KB. Those disks could report out of space with only 50% data on it or even less.

It doesn't happen on tape. There is no empty space on a tape. You can therefore backup a 2 GB disk on a 1 GB tape.

top | Art Quarter | feedback

Q. Should I save my internet downloads to tape?

I would not do it because of access technique on the tape. This was state of the art with them old Commodore CPM machines used by kids. They stored programs and data on audio cassettes. But that was a different world. Store them on disk, and whatever you don't need for fast access, store that on tape. If you need it, restore from tape to disk.

There are different modes of tape backup. If you choose an image backup, you get all stored like on disk which is fast for reading, but, fragmentation in mind, it is clear that this can be awful if not impossible to restore single files. It is best to restore the disk as is, when you want to replace the disk, for example. But then you might have a different model, so image is not feasible either.

Therefore you will backup almost always not in image mode. If you backup a session on tape, then all files will line up fine, so restore does not take that long.

But for easy access and backup purposes, there is another technique these days. Why not record it on CD? Mostly you can't alter data once burnt, but this is possible, too, although at more cost (rewritable technique).

top | Art Quarter | feedback

Q. How to deliver content.

Common printers not being able to reproduce pictures as lively as monitors, it makes no sense to print them. On the other hand, it makes no sense cutting off a delicate and intricate topic. Therefore I do not care about length. If things get too lengthy, I split to two or more issues.

Short articles are ok, but then... Corey Rudl advocates strongly long copy, and I think he is right. You probably know this saying:

How does the minister deliver his mission on Sunday?
1. He announces what he is going to say.
2. He delivers his message.
3. He recalls his message.

Put more funny:
1. He says what he will say.
2. He says it.
3. He says what he said.

If you put it in signal/noise ratio, that's a lot of noise. But we all know, 80% of a message has to be noise to be comprehensible at all, kind of a lubricant. Just listen to a normal talk - most of it will be small talk. The really important things happen in between.

I notice when I skim my mail, I get interested after just a few words. I gather content, and content is not to be separated from form. Content can't be delivered in just a few words, so it is form that does signal there is content to come, so it will pay to stay tuned and read on.

The more we are bombarded with information, the more form will become important, as form sells content. See web design for example. Hence I think ezine design will become more important, and soon people will find out that ASCII mode just can't get you as far as you want to go.

This argument is not to defend the style of my ezines, however. I can't go without pictures, so I have to use html and take care of the hassle that goes with it (right now I got problems with my service having a bug with html). But of course, I have to use form to deliver content, too.

All this discussion does not apply to ezines only. I once read an essay of a German novelist, Arno Schmidt. Subject: How do fellow novelists capture the reader? Investigation of the first words, first sentence, at most first paragraph only of diverse famous novels.

I recall a German ghost speech writer saying: "Man kann über alles sprechen, aber nicht über 20 Minuten." I can't translate the pun, so it reads awkward in English: "You can talk about everything, but not more than 20 minutes." This is simply not true.

I still remember vividly a speech of pretty exactly 120 minutes (the goal of the orator, he checked often to keep it) without any pause in front of an audience of some 800 with only a microphone, no water, no lectern, no manuscript, no media, and, what's most astonishing, without any one of the 800+ audience missing one single word.

According to popular scientific theories this is just not possible. But I could tell you for half an hour about that lecture, and I bet most of the audience would be able to. This person wrote a book (not only one, many). And he delivered this speech as part of a promotion tour for that book. Or rather, for what he had to say, which he could not even put into one book, let alone one speech. What keeps the reader or audience interested and awake is signal wrapped up with a nice mixture of noise.

Now, if things can get really long, what can you do? Alternately to sending the html mail I offer a short text only URL announcement for the online archive. Online time may be expensive, so I suggested a download tool for offline browsing. After a while, I only mentioned it, but nobody asked me for the address so far. I personally can't do without. Whenever I stumble upon something interesting which takes more time, I just download the whole site to my disk for offline reading when appropriate.

top | Art Quarter | feedback

Q. How to use email links.

If you gather email addresses, you will experience lots of bad addresses. They cost you time and effort. Worst thing I witnessed was someone on instead of or vice versa.

I used to query the address via html form input. They had to supply their address and this was prone to error. I see this all the time and can't understand why people do it that way. I've seen confirmation messages and such stuff: "Is this really your address? Double check ...."

Heck, I'm a surfer like anybody else, this is just not what I want!

Therefore, after having followed the main path first and experienced the unavoidable troubles in consequence and then tried to implement such an "I appreciate your time and effort, but please do me one more favor and check ..." message (which was not easy) I dropped the whole thing.

Instead I switched to mailto links. Just two clicks with the mouse. That's it. Convenient for the user. No chance for misspelling on the visitor's side. Still some invalid addresses. Well, if they put the wrong address in their mail program...

I see this in lists often, so here's the how to:

Syntax in text mail:






I use it to track my ads. Subject=WAHP or whatever gets into my mind will tell me this is an inquiry from my ad in "Work at Home Parents". I can use it for a filter, too.

Don't use blanks in text mail, it won't work any more. I see this all the time. Caveat: There are mail programs not supporting this syntax, but anyway, everybody understands what you mean and can do it manually.

If you use this syntax with html mail, it will not work (well, life ain't easy). Html mail and html web page syntax are the same, just a normal link syntax (Mail me, please!). Here you can use blanks. See Wanda's site for an example (I copied from Wanda). Here, too, are browsers not supporting this technique. If you want to exclude this risk, repeat your syntax in the link text or somewhere on the page (see for example Pablo 1.5).

I even based my order system on this technique. To me as customer, all those shopping carts and CC forms are just tedious. Mail links are fast and easy (although they do look unprofessional - might be a strong objection).

This is a strong objection to OneList, too. When I get an offer to subscribe to a OneList list, I know I have to go online, log in, supply my user name and password, and there are still 2 clicks to get listed. Lots of trouble for a benefit I am not sure I will appreciate in the first place.

No trouble with them clever guys/irls using the correct mailto syntax and a service based on simple mail subscription. That's why I have my lists with .. err, no, I won't interfere with your biz, Nancy.

Bottom line: Mail is used for all sorts of things. With proper syntax, you make it a breeze for you and your customers.

Proper syntax for web pages in text mail:

Don't forget the http and the : and the //!!!
:: and /// won't work either!! No blanks!

Well, just another tip:

If you have an index.html or index.htm or default.htm or default.html in every directory on your site, you can point your visitor to your specific page with a professional looking address:

(no index.html or so necessary; AOL won't allow you to create 3 subdirectory levels, I guess, but I don't know. Ask your ISP if there is a limit. Geocities has 2, AFAIK)

Julie Groat wrote in reply to this:
> You know what one of my pet peeves is? People who don't have the
> documentation to back up what they claim.

She warned that older mail programs would crash with this syntax. My reply:

I'm sorry. But... I'm thinking about how progress is achieved. Certainly not by waiting them old programs being redesigned and customers all being upgraded. You can wait long for that. People just won't upgrade at all if they are fine. And they are happy if their machine runs without trouble. I act so myself. Static structures tend to be stable. No progress at all.

I got different advice. It is not my invention, though. It is just what happens all the time, if I want it or not. All I can do is swim with the trend and learn a little on my way. So here it is:

Crash 'em as you can! Until they will switch to a program that does not. Which will in turn force the old program owner to fix his problems. And invent some more nice things we all need but don't know about yet. Which in turn will persuade some to use the new features. Which in turn will crash the other programs. And so on.

top | Art Quarter | feedback

Q. Experiences with dictation systems.

Preface after the fact: I started out small, but lots of emotions and remembrances came up. This system and experience gave rise to so much thought I obviously wanted to share with you. So please take my apologies, maybe there is some grain of gold in it for you, too.

I went with Dragon Dictate back in 96 when they were on DOS. We as dealers had to spend $6,000, price was $9,000 plus taxes.

I believed in the system. It was hard, but I tried all I could. Then they switched to Windows. IBM came up with their system. I tried it, too. Then they improved. We tried to sell it another time. Things were really easy. Then came Naturally Speaking.

Well, here I sit and type. I installed Naturally Speaking and saw it worked fine. I had a 90 MHz chip and they wanted 200, I had 32 MB and they wanted ... So you could suspect it would work with the appropriate machine.

Now I got this 200 MHz 64 MB machine and have the disc ready to install. I don't. I'm sick of the hassle. I spent so much time trying to get this thing going. I rather type along.

We tried to sell it to lawyers. They have to pay typists, depend on them, so we thought it would make sense to them. There are some clients using it. Trying hard to. But I asked myself: Is there any sense getting them into that?

Lawyers are trained to think and invent their "copy". They are extremely fast dictating to a machine. Typists are extremely fast typing from a machine. Of course they could do more interesting work, too, would be more gratifying also, but then...

You can achieve much with standardizing and using form letters. They could work this out much more instead of wrangling with this technical system. Would probably free their time and the time of their typists much more than using this dictation system.

I made a test with a moderate fast typist. I dictated the same 3 paragraphs several times to give the system a chance. I dictated life. She was faster.

Now I am not a lawyer, I am not a typical dictating person, in fact I tried but could not get into that. So I am sitting at my keyboard doing it myself (I guess you all, too). I am quite fast, but not very good. In fact, I learned it myself which is not good to begin with.

This is dumb, I know, I should dictate this letter to a machine and have a typist write it. But I got none and never had one, even when I had 10+ employees. I am not used to it. Needs some kind of discipline, I guess.

So I had reasons to use this machine, much more than my clients. I was extremely enthusiastic about it. But I had so much troubles with it, I spent so much energy and emotions with this thing, I felt like an alpha tester be abused over and over (I guess you know this feeling in this kind of industry).

They will succeed one day. I realize they need people like myself being naive enough to get this thing going in the first place. Drs. Baker will be famous for centuries for their work, that's for sure. This was the key question of my customers: When is the right time to jump on that bandwagon? I don't know.

The experiences of Ted Brown are mine. He put it technically, I personally. A technical tip:

> 7. Personlly, I found the headset annoying.

Obviously, this was typed. A dictation system makes no spelling errors (except you let it learn them, which is another topic: How to work with these systems). I found that the headset they recommend and deliver and make a lot of fuss about is not only annoying but also of bad quality!!!

This I cannot understand at all. I had my best results with the desk microphone delivered with my Soundblaster AWE 32 sound card. This thing must be cheap as it is bundled free with the card. I got my hands free and recognition was far superior to the headset. As a reseller, I had direct contact to their technicians, but I never got an answer to this issue. The salespeople did not work with it anyway. They tried to sell not by example but by copy.

IBM had a different approach, similar results, but I found Dragon better for daily use. Naturally Speaking is a blend of the two. Hopefully it is a junction of the merits and a exclusion of the negative aspects of both systems, but I can't really tell.

If I would work with it again after overcoming all installation troubles, I'd probably agree with Ted:

> 9. In the final analysis, taking into account the training time and other
> initial hassles, I would not be without it when I have significant text
> to enter.

Emphasis is on "significant text".

Of course, I got the German edition. Finding myself writing English all the time now, I would have to buy the English version. I'd certainly hesitate to do that.

Maybe the cheap version K. Williams talks about will do (Dragon Simply Speaking). I'd try it. Maybe I will myself. My version needs a dongle so it could by no means work by downloading the program. But software does not need a dongle by default, of course.

I looked at this experience from a point of product development, too. I saw it with other products, i.e. Novell. First proprietary hardware, then hardware copy protection, then software copy protection. Hardware and software development needs enthusiasts to take place.

We would not have sound cards if people were not enthusiastic about playing games on the computer. Same with video cards. People do spend money on games, not much, but sufficient to boost the sound and video industry.

Text recognition is an old hat. Reaches back to the 60ies. But it is no fun thing. People would not spend money on it. How to get this thing started? They worked with government and military funds. Sold at high price when they got a version that could possibly be offered (point... shoot... aim!!!).

The Bakers always stressed the fact they were privately funded. At some time Seagate was welcome to join with money. They don't tell you, though. IBM did it right away with a big purse. (Bakers started in their lab.)

I think this is an achievement in itself, to have got this thing so far at all. If not for the Bakers, IBM would not have got it by themselves. Hail competition and private enterprise!

I thought they should have tried it for small money and mass market, to get all the players involved. IBM finally cut prices so hard, you can't believe it. When we set out our big campaign back in 96 (cost us a lot of money, but we wanted to make a lot with this system, too), first question was about falling prices.

I reassured the customers, having been concerned myself. My research showed they had patents and were way ahead of competition. They had invested so much they could not afford to lower prices. They would not have to.

I was wrong. In the midst of our campaign they cut prices by 50%! I got the message on the road driving from the morning show to the afternoon show. The day before I had sold two systems to a single office for $18,000. Imagine their call when they found out. It didn't take them a week.

3 months later they cut another time by 30%. A colleague of mine had ordered for 100,000$ to become an A type reseller. You can figure out his loss. IBM drove its own product to the $99 level and then integrated it into OS/2 at no cost at all, even promoted it as a main feature. Nobody talks about OS/2 anymore, let alone about their voice system.

My guess is: It will be as natural as talking to a dictating machine in some time in the future. Everybody will be using it, nobody will use a keyboard or a mouse. Until then, we'll stick to our old equipment. One more example to illustrate a universal law:

I don't use a mouse for years. I use a pen. I was doing missionary work on that device. I think the inventor of the mouse must be rewarded the prize for the worst instrument invented ever to do a specific type of work. But I could convince only my programmers. They loved it. The other employees would only try if I told them so and oppose nevertheless, so I let them their mouse.

I can use a mouse if I have to, but I hate it. I'm much more productive with a pen. But I realize people stick a long time to what they used always, what everybody else uses, what they know. We have a saying here: "What the boor doesn't know, he won't stuff with."

Same with scientific truth, by the way. The new builds up in the background, very much opposed, until critical mass is reached, then it behaves like the old system very soon, hampering development.

Communities... There are people (Hellinger for example) finding satisfaction in dropping a sentence in communities, watching it being very much opposed to, to witness somebody else repeating it the year later as own insight, being acclaimed at unanimously.

PS: I started using Dragon Dictate again as my fingers hurt. RSI Syndrome (repetitive stress injury). Works fine.

My post to MarkeThink 01-10-99:

Hi folks, I'm glad to be "on-line" again. Ever since John moved to a listserver, the digest bounced back, so I had to read the archive on the web later. Happily, the problems seem to be resolved now.

I remember, that I reported about using DragonDictate because of RSI syndrome. It is time to give another report for all those of you working hard and long at the keyboard.

My DragonDictate was a German version and I had to abuse it to accept English pronunciation and vocabulary. I told you about Dragon NaturallySpeaking, Jim Daniels announced to use it, but he never mentioned it again.

Well, after some time it happened that I had to dictate something in German. I owned DragonDictate for quite some time, but I didn't have the opportunity to use it on my faster machine (200 MHz vs. 90 MHz, 64 MB RAM vs. 32 MB RAM).

I was amazed how fast and precise the German version worked. I figured that this version was optimized to German phonems. This should be the reason why it didn't work as good with my tweaked English setting. I decided to buy the English version of NaturallySpeaking. As I told you back then, it is not expensive at all.

Remember, I started working with this technology (trying to sell, too) back in 1993. I knew the IBM system as well. Set up and training was fun, but the first attempts to dictate weren't as successful as expected. So I kept on dictating, and after some 10 minutes, it showed more and more success. I mean, this system worked better than my tweaked German DragonDictate.

Nevertheless, I was puzzled about some of the problems I encountered. As you may guess, I found out. Recently, I had the pleasure to try out a new service I offer on my pages (at bottom), a free call. I asked Wanda Loskot to call me, and indeed this service worked extremely fine. The first thing Wanda noticed was my British accent! (See her at

Now you see, my teacher at high school was British. I tend to pick up accents wherever I live, so it's no wonder my spoken English is mixture of German, British and southern American accents (30 years ago, I spent a year in New Orleans).

Dragon NaturallySpeaking comes in two flavors: British and American English. I installed American English because of spelling rules (having witnessed American sensitivity to foreigners), but after a while it occurred to me that I could improve on spelling errors if I tried to pronounce more the American way.

Well, this proved true. It is not easy for me to pronounce like Americans (remember, I didn't have much opportunities to write or speak until I entered the Internet), but if I manage to, the system really works great. I like to cite Jerry Pournelle: "highly recommended".

Having found out this, I thought my German NaturallySpeaking version should work really good as I don't have any problems staying within my peculiar accent (everybody has some kind of accent, but only foreigners can have a mixture of different accents).

And in fact the system works even better with this "native" version. Bottom line: give it a try! (No, I am in no way affiliated with the producer and won't get any commissions or rewards for telling you this).

PS: I think I ought to tell you one caveat: you need to discipline yourself to use the system in each and every instance. It is really easy to switch to the keyboard just to type in one or two sentences as a reply to an email.

This is a big mistake. Not only do you spend significantly more time typing, you also have to cope with spelling errors and eventually ruin your hands in addition.

I hear that they have systems for people who never type themselves but dictate to a machine for typing by a secretary. In this case, you dictate to a digital device which in turn feeds Dragon NaturallySpeaking. As I don't have a secretary, I can't tell you how good this works.

top | Art Quarter | feedback

© Copyright 1998
Texts and Images

Galleries | Artists | Journals | About | Mission | Site Map

© 1998,1999 · Werner Stürenburg · Germany  ;· Hauptstr. 13 · 32609 Hüllhorst · Tel. 0(049)-5744-511-574 · Fax -575

Read my Blog · See also my personal site