Many users access online services by calling them directly. A lot pay extra for long distance calls to other cities and countries, even when this means inferior transmission quality (like when noise characters degrade the data). Others investigate other routings for their data. One option is the packet data networks. Most countries have Public Data Networks (PDNs) operated by local telecommunications authorities. These services are often cheaper than direct calls for some applications, but more expensive for others. Before using a packet data network, you'll need to establish a "Network User Identification" (NUI) with the PDN carrier. You must also know the Network User Address (NUA) of the hosts that you want to access. In Scandinavia, the local PDNs are called Datapak. They can be accessed by direct local calls or through leased lines. To personal users, direct calls are least expensive. A leased line may be cost efficient when the daily volume is high, like in a company. When you communicate with online services through a PDN, the latter will split your data and bundle it in standard envelopes or 'packets'. Each packet is marked with a code and sent out into the data stream. Based on this code, the packet is routed from computer center to computer center until it reaches its final destination. There, the information will be reassembled into its original form before being handed over to a user or online service. It is almost like traveling by train. The price per packet or traveler is lower than what it costs to rent the whole train for your trip. National telecommunications monopolies were the first to offer packet data services. Their rates were moderately lower than for long distance calls, but it was hard to find the relationship between real costs and prices. This is still the situation in many countries. Throughout the world, efforts to privatize nationwide phone networks continue. In many countries, this has given us some interesting competitors offering attractive rates for similar services. Their rates differ considerably from country to country, as does the quality of transmissions. The advantage of using packet data also varies considerably, by application and by country. The best routing for retrieval of online news may be impossibly expensive for chats or complex online jobs. We can offer no hard rules of thumb, except this:
Compare rates regularly!
What is cheapest? ————————- Some networks charge by the hour, while others charge by volume (number of characters transferred per minute). When volume is low, your best bet is to use network services with a low price per minute and high prices for volume. When volume is high, you may be better off using those charging by the minute. To estimate costs reliably, you'll need statistics. Since your usage probably differs from what others do, start accumulating experience data now. Like this:
On services only charging for connect time—————————————————————Capture trip information to a log file. Register the followinginformation:* number of minutes connected* modem speed* number of characters transmitted.Some communication programs can do this automatically for you.
On services charging for time and volume————————————————————Log the following information:* number of minutes connected* modem speed* number of segments or packets (measurements of volume)
You need this to estimate the average volume of data transferred by minute. Here are some general experiences and hints: Long streams of data without stops are cheaper through services that only charge by the minute. Retrieving software is a typical high volume application. Trips that include navigation from conference to conference, with a little bit of up- and downloading here and there, make the average transfer speed fall dramatically. It's like driving through a big city at 150 kilometers per hour. Red lights will reduce the average speed considerably. The actual transferred volume of text per minute will differ from place to place (geographically), and often also from call to call. It depends on factors like:
* How fast you can enter commands and how much time you spendstaring at the display before pressing keys,* How long it takes for an online service to react to yourcommands. For example, the response time on CompuServe at04:00 GMT on a Friday morning (it is evening in the U.S.) ismuch worse than at 10:30 GMT on a Sunday morning. Then, mostusers are asleep.* The load on your packet data network while you use theservice (or the amount of noise and retransmission, whencalling direct),* The type of modem you are using (speed, level of MNP),* The number of commands you (or your scripts) have to enterduring your online visit. An increase in the number ofcommands, reduces the average transfer speed.* The amount of transfer overhead for color and screen handling(like, VT-100 codes) that is transferred with your text.* Your use of menus and help texts while online, or whetheryou come as "expert" with a minimum of prompts.
It's impossible to calculate the practical effects of these items. You will just have to bear them in mind when estimating typical jobs, measuring speeds, calculating costs, and comparing networks. Finding the optimal network for our needs, will take time, but is well worth the effort. I think the figures may surprise you. The network services in this chapter will often give you better quality transfers than a direct call. On the other hand, calling direct may give more characters transferred per minute. The average speed tends to drop dramatically when using a packet data service.
Using national packet data services —————————————————- Most commercial online services can be reached through national PDNs, but you may have problems finding the correct NUA (Network User Address) to get there. Few PDNs have a directory of available "electronic telephone numbers" for you to consult. The Norwegian PDN, Datapak, used to be my only alternative for access to foreign online services, and I thought that the cost was acceptable. Not so anymore. My applications require that data be pumped back and forth at maximum speed. On network services charging by a combination of volume and time, 80 percent of my costs are typically for volume, while 20 percent is for connect time. When I log out after a successful visit to CompuServe through Datapak, the two services give me similar reports:
Thank you for using CompuServe!
Off at 10:11 EST 24-Nov-87Connect time = 0:15
The last line comes from Datapak. It tells that I have received 537 segments and sent 75. The "Segment" is Datapak's volume measure. A segment contains up to sixty-four characters and/or carriage returns. The price is calculated accordingly. At today's prices, Datapak is still my cheapest alternative calling CompuServe for chats. I use Datapak when connecting to TWICS in Tokyo, as the only alternative today is direct calls at a prohibitive cost. Once i-Com (see below) starts offering outdial to Japan, I expect this service to be substantially cheaper. The slower your modem speed, the more attractive is Datapak compared with direct calls. To get access to a national PDN, you must have a user identification and a password. (Getting temporary access to PDN services while traveling abroad is often hard and expensive.)
| Note: If you have access to a national PDN, but need | | information about PDNs in other countries, try Hostess, the | | Global Network Service's information service from British | | Telecom in England. The NUA is 02342 1920101013 (02342 is | | the Data Network Identifier Code section of the address.) | | Username or password is not required to use this service. |
Outdial through PC Pursuit ————————————— Sprintnet (formerly GTE Telenet) lets American users call bulletin boards in North America at lower rates through their PC Pursuit service. They pay a modest subscription to call a local number for access to PC Pursuit. Once connected, they can enter an electronic phone-number to connect to a so-called 'outdial modem' in another city. Once connected to the outdial modem, they can give it dialing commands and have it call any local number. This way, they can use PC Pursuit to call an online service in the area, or the private modem of a friend. We call PC Pursuit an Outdial service. Such services normally offer lower rates for access to remote bulletin boards than what it costs to call by long distance. Besides, they reduce the chances for noise on the line.
Outdial through i-Com ——————————- i-Com offers outdial to North American online services by reselling capacity from Tymnet's network (owned by British Telecommunications PLC). In the United States, Galaxy Telecomm Corp. offers a similar service under the name Starlink. Outdial to numbers in Japan and Europe is planned. i-Com markets its services to users in Europe and Japan, and have local access in Brussels, Paris, Lyon, Milano, the Hague, Eindhoven, Zurich, Geneva, London, Belfast, Birmingham, Bristol, Cambridge, Edinburgh, Leeds, Frankfurt, Cologne, Munich, Madrid, Stockholm, Copenhagen, and more. The basic fee for access to the service is US$25.00 per hour (1992). You don't pay volume charges. The monthly subscription fee is US$8.00. You can pay using VISA or MasterCard/Eurocard. In Norway, I have used i-Com to connect to The Well in San Francisco, MetaNet in Virginia, EXEC-PC in Wisconsin, and SciLink in Toronto, Canada. At the time, i-Com was cheaper than direct calls and Datapak for access to these services. While an ID on your local PDN is only valid in your area or country, your i-Com ID can be used all over the world including several cities in North America. Once your plane has landed in Milano, you can dial the local i-Com node to connect to your favorite service. i-Com also has a bulletin board (US$13.00/hour). These are some of its services:
* Search a database to find BBS numbers in a given area ofinterest, or to locate outdial numbers in a given city orarea code.* Conferences about how to use North American bulletin boards.* Retrieval of shareware and public domain software.* Online shopping of American goods at American prices.
Cheaper access to CompuServe —————————————— Wherever CompuServe has local access points, you'll be better off using these. You do not have to sign any special agreements. Your CompuServe ID is all you need. Payment for using these services will appear on your CompuServe bill. CompuServe has special deals with a list of network services, like InfoNet Europe (formerly Computer Sciences Corp.), Istel, FALNET, FENICS, CompuPass, LATA Networks, Tymnet/Sprintnet. Enter the command GO LOG on CompuServe to get access information, and GO RATES for rates. I have been using CompuPass from Japan, CompuServe's own network in the United States, Istel, InfoNet, and PDN services in Europe. When at home, I usually use CompuServe's 9600 bps node in Stockholm, Sweden. It is even cheaper than calling Oslo for a 2400 bps node for most of my jobs. There is no surcharge when accessing at non-prime time, and US$7.70 per hour during prime time (weekdays 08:00 to 19:00 local time). In addition, I pay long distance rates to call the node. CompuServe has no extra charges for volume.
| Whenever CompuServe opens a new node in your vicinity, or | | upgrades the modem speed on one of their nodes, look at the | | effects on your total costs. | | | | Use software for automatic access and navigation (like TAPCIS,| | OzCIS, or ATO). They give higher volume per minute and make | | your accesses even more cost efficient. |
Before leaving for a business trip, visit CompuServe to find local access numbers in your destination cities. The list of countries includes Australia, Belgium, Canada, Denmark, Finland, France, Germany, Hong Kong, Italy, Japan, Mexico, Holland, South Africa, Spain, Sweden, Switzerland, and England. You can also access CompuServe through i-Com and other outdial services. CompuServe has exchange of electronic mail with Internet. You can also access the service by telnet to hermes.merit.edu (binary transfers are impossible, though).
IXI - a European alternative to PDN —————————————————- IXI is an X.25 data network for European academic, industrial and governmental research centers. It is sponsored by the EEC under the ESRIN project, and is operated by the Dutch PTT. IXI interconnects national research networks, many national public data networks and several specialized international networks. It works like a national PDN service, but uses its own Network User Addresses. Echo, STN, DIMDI, Data-Star and other database vendors can be accessed through IXI. The service is not available to most users having email access through the Internet.
Using DASnet to cross network boundaries ———————————————————— DA Systems forwards electronic mail and files (also binary files) across mailbox system boundaries for customers. They can send your mail to several large in-house systems, information networks, and over 60 commercial mailbox systems in 30 countries. These are some systems on their list: ABA/net, Alternex (Brazil), ATT Mail, BIX, BITNET, CESAC (Italy), CIGnet, ComNet (Switzerland), CONNECT, Dialcom, Deutsche Mailbox, Dialcom, Envoy 100, EIES, EasyLink, Euromail (Germany), FredsNaetet (Sweden), Galaxy, GeoNet (hosts in Germany, England, U.S.A.), GreenNet, INET, INFOTAP (Luxembourg), Mailbox Benelux, MCI Mail, MercanMail (Asia), MBK Mediabox (Germany), MetaNet, Nicarao (Nicaragua), NWI, OTC PeaceNet/EcoNet, Pegasus (Australia), PINET, Portal, PsychNet, San Francisco/Moscow Teleport, Telexphone (France), TeleRede (Portugal), Telehaus Nordhorn (Germany), Telemail, TEXTEL (the Caribbean), TWICS (Japan), UNISON, UUCP, Web (Canada), The WELL, Internet. This list may suggest lack of connectivity between networks that do indeed have connections. For example, Internet email may easily be sent to ATT Mail, Alternex, BIX, BITNET, FredsNaetet, GeoNet, GreenNet, and many others on this list. Connectivity changes constantly. Check to see if you really need it, as this service is far from free. DASnet also lets you send email as telex, fax and by ordinary mail. They charge you by the number of characters transferred, and the destination address. (Contact Anna B. Lange, DA Systems, Inc., U.S.A. Tel.: +1-408-559-7434, or write her at AnnaB@11.DAS.NET).
FidoNet - grassroots playground ———————————————- FidoNet is an amateur network consisting of tens of thousands of bulletin boards all over the world. The network is "loosely coupled," meaning that most of the participating boards are not always connected. They call each others at regular intervals to exchange mail, often in the middle of the night when the rates are low. Most FidoNet boards have conferences, and allow you to send mail to users of other systems. NetMail is a term often used for private FidoNet email. EchoMail is used about its international conferences. The selection of echomail conferences on a given FidoNet board can be as unique as the rest of the system.
RelayNet ———— is another global network of bulletin boards. It offers exchange of email between systems. Messages and conference items entered on one system will automatically be copied to other participating boards. Your costs for "talking" with others in other parts of the world are very small.
Other grassroots networks ————————————- It doesn't take much to set up a bulletin board service, and it is as easy to connect BBS systems to each other in a dial-up network for regular exchanges of email, files and conferences. All over the world, grassroots networks keep popping up with names like ILINK, AmNet, Suedd MB-Verbund, Starmail, MagicNet, A- NET, MausNet, Zerberus-Netz, SMBX-NET, BASA-NETZ, you name it. Many boards offer access to more than one grassroots network, as well as to the Internet. Thus, the ability to send global email is extended to new users every day.
Other services———————The PDN Connect-USA competes with Starlink in North America.(Connect-USA Communications, Inc., 2625 Pennsylvania NE Suite 225,Albuquerque, New Mexico. 505-881-6988 (voice), 505-881-2756 (FAX),505-881-6964 (BBS). )Global Access is a similar service reselling time on theSprintnet network in North America.
Reducing the cost of using mailing lists ———————————————————— The problem of subscribing to mailing lists is that all discussion items come to you in individual messages. Each message comes with its own mailer header, and this information is generally completely useless. (Read "Returned mail" in Chapter 7 for details.) Newer versions of the BITNET LISTSERV software provide commands that solve this problem:
SET DIGEST
———————————
This command is sent to a LISTSERV to make all daily messages
come to you in one, single message. Example: Say you've joined
KIDCAFE@vm1.nodak.edu, which usually has a large number of
messages each day. Send the following command to the LISTSERV:
It will typically reduce the number of lines received from this mailing list by around 50 percent.
SET INDEX
——————————-
This command is sent to a LISTSERV to get a daily list of
messages, like in this example from KIDCAFE:
Index Date Size Poster and subject——- —— —— —————————22839 06/22 26 From: David Chalmers
Based on this list, you can use the LISTSERV's search commandsto retrieve individual messages of interest. These commands aresimilar to those used for searching in chapter 7.
For more about searching mailing lists' message bases, send a message to LISTSERV@vm1.nodak.edu with the following command in the text of your mail:
Some LISTSERVs offers simplified search commands and macrosto make retrieval of individual messages simpler.
Little is gained by being skillful atlocating and accumulating information,and then becoming drownedin an avalanche of datathat one cannot manage - or use.
This chapter starts with how to build a personal data base on your own hard disk. We continue by investigating strategies for finding interesting information on your disk, before winding down with some words about what separates good information from bad.
Search and throw away ——————————- To novices, everything is difficult. During the first online trips, they may feel as if moving to the other side of the globe to start over: They need new newspapers, magazines, information sources, and services. Trial and error are required to find online gold mines. As you get more experience, focus tends to shift from getting information to digesting. Getting the data gets 'into your fingers', and doesn't bother much anymore. The number of retrieved lines increases. The only bad news is that your reading speed remains at the same old level. In our time, people tend to talk more than they listen, and you usually find more information than knowledge. Therefore, say NO to irrelevant information. It is seldom worth keeping. There is generally no good reason to learn things that you really don't need to know. Practice "selective ignorance." Regularly evaluate your online sources critically, and discard those costing you more than they are worth. Concentrate on those giving the best returns. Adjust the frequency with which you visit selected services to match their usefulness. What used to be daily visits, may have to be downgraded to once per week or month. Consider replacing daily news monitoring by clipping services. Plan 'overview' and 'details' with different frequencies. 'Overview' refers to online trips to get an impression of what generally goes on. An example:
My script system is set for automatic visits to the CompuServe Toshiba forum. Whenever I visit, it 'digs out' unread messages based on key words on the item's subject line. During 1991, it searched for these strings: '5100', T2000', and '425'. Once, This gave the following message to read:
#: 29550 S6/Hi-Power Notebooks05-Oct-91 17:27:30Sb: #T2000SX RechargerFm: Steve Kitahata 75166,1741To: All
I tried to order the battery recharger for my T2000SX from Jade Computer last weekend. The sales rep said it would take about a week, so I called today to check up on it. He told me that I could only buy the recharger with the car adapter as a bundled set for $260. They had both advertised in their flyer as separate items, which they should be. Has anyone heard of this?
Does anyone know of any sources that have the battery recharger available? Any help would be appreciated. Thanx.
— Steve
My script found the search word "T2000" in the subject line's text (Sb: #T2000SX Recharger), and subsequently selected the message. Once per month, the same system "scans the horizon" to give me an idea of what is going on. This is done by requesting a list of subjects being discussed. Here is part of one such list:
29555: DOS 5 Upgrade6 replies
29540: TDOS Upgrade questions3 replies
29585: Toshiba DOS 5.0 ships!1 reply
29586: DOS 5.0 Upgrade Solution
29580: ToshibaDOS=bad business8 replies
29581: DOS 5 / Stacker1 reply
Reading the list, allows me to see if new and interesting topics are up for discussion. If I use Stacker and want contact with other users, I can request message number 29581 and the subsequent reply (1 reply). That should give me some email addresses.
| Several advanced communication programs and offline readers | | have built-in quick scan features. For example, TAPCIS does | | this just fine in CompuServe forums. | | | | When retrieving conference messages from bulletin boards using | | 1stReader at high speed, like 9600 bps or above, then the cost | | of downloading all new items may be insignificant. Therefore, | | you might just as well do it. | | | | Later, when reading the captured mail, 1stReader lets you | | select messages to read from a list of subjects. You can save | | what you want to keep, and delete the rest. |
By regular scanning subject headers you reduce the risk of missing important trends, for example because authors were using other terms on the subject line than expected. Scanning also lets us discover if the discussion is heading off in other, interesting directions. After a while, you'll have a set of sources, persons, and tools that will provide you with what you need. This is your personal infrastructure of electronic information. Now, you must maintain and cultivate it.
Store incoming information ————————————— Chances are that you will retrieve more information than you can read. Sometimes it takes weeks for me to get up to date with my own readings. If you visit several online services, consider storing the data in files with different names. Use part of the file name to show the source of this information. If visiting a service regularly, consider using the date as part of the file names. This will make it easier to select, read and search them in a useful sequence.
| Example: Say you're regularly visiting TWICS in Tokyo. What you | | download on November 10, you may store in a file named | | | | TW1110.TMP | | | | My scripts do this automatically. On some services, they also | | split retrieved data into URGENT and MAY BE READ LATER files. | | Private mail from TWICS is stored in NB1110.TMP. By storing | | private mail separately, it is easier to see if somebody wants | | a quick reply. |
All file names in this example have the extension .TMP (temporary). This signifies that these files are unread. When I read them, and select parts for permanent storage on my hard disk, I use different names. Often, I use the year, or a month/year code in the file name extension. For example, the file DIALOG.93 contains information from DIALOG collected during 1993.
Postprocessing the data ———————————— The data capture is completed, and the retrieved data is stored on the hard disk in more or several files. Your next task is to
* Read the received texts,* Cut and paste selected parts to archive or work files,* Prepare responses to your electronic mail. This may includequoting part of the incoming messages in your replies.* Finally, delete all temporary files.
Many advanced programs have these features built in. If not, you may use your favorite word processor, or something else. There are many alternatives. LIST is my favorite MS-DOS shareware file viewer program. It can be downloaded from most bulletin boards. Using LIST, it is difficult to destroy your precious retrieved data while reading, cutting and pasting.
| MORE ABOUT LIST: | | Assume that all input data is stored in the disk catalog C:\IN | | and that you're using the file name convention suggested above. | | Type LIST and press Enter. A list of file names will appear on | | your screen. Press S to sort the list, and then D to have them | | sorted by creation date. The newest files are at the bottom of | | the list. | | Move the cursor (using the Arrow keys) to the input file | | that you want to read and press Enter. Scroll up and down in the| | file by pressing the PgUp/PgDn or the arrow keys. | | Let's assume that you are reading TW1110.TMP right now. | | On your screen is a piece of information that you want to | | keep for future reference. Mark the text with ALT-M commands | | (keep the ALT key pressed down, while pressing M), and then | | ALT-D. LIST will ask you for a file name. You enter TWICS.93, | | and the text is appended to what is already there. | | This method allows you quickly to mark and append parts | | of your input file to various archive files. Press ESC to | | return to the file list when through, then press D. LIST asks | | if you really want to delete the file. Press Y, and TW1110.TMP | | is gone. | | LIST lets you find information stored in your archives | | (string search). What you find can be marked and copied to a | | work file. It can also be set to invoke an editor or a word | | processor for the selected file. |
Reuse of data on your hard disk ———————————————- Over time your personal archives will grow in size. You begin to experience the benefits of having all this information on your hard disk. Yesterday's news is today's history, and may be used in many interesting ways. One business executive regularly monitors key technologies, customers, competitors, and suppliers. He does it by tapping sources like KOMPASS, Associated Press, and Reuters. Interesting bits of information are regularly stored on his disk. Tomorrow, there is an important meeting with a major customer. First, a quick search through the personal customer database to be reminded of important events since the last meeting. An unfamiliar person is also going to be present. Maybe there is some background information, for example about a recent promotion. Then, a quick check on major competitors. Maybe they are up to something that he needs to know about. With efficient tools for searching your hard disk, finding information takes only a few seconds. If you are still left with open questions, go online to complement. On MS-DOS computers, you can search the files with WordPerfect, LIST, the DOS utility FIND, and a long list of other programs. I prefer programs that let me search for more than one word at the time, like in HYDRO AND PETROCHEMICAL AND CONTRACT, or EXXON OR MOBIL.
| MY FAVORITE: My favorite search utility is LOOKFOR. It can | | be downloaded from many bulletin boards. The MS-DOS program | | is small, fast, and is superior for searches in DOS text files.| | Store your finds in work files, or print them out on paper. | | LOOKFOR is not an indexing program. It is ready to search | | anywhere, anytime. |
Discipline and organization is required to get the most out of your file archives. You must decide what to do with each piece of information: Should it be printed out and be read in front of the fireplace this evening, or should it be circulated? Should it be stored on your hard disk, or be refined before storage? Use standard file names that are easy to remember. If you don't, risk having to view files to find out what they contain. It may take longer to find a piece of information in a casual file on a large disk, than look up a piece of information on paper in your inbox. Therefore, finish handling your capture file while you read it on your screen: Send the pieces to their final destination. Make immediate transfers to your TO-DO files. Give the original file a name that makes it easier to move later. Have a procedure that prevents duplication of effort.
Desinformation, deception and errors —————————————————— Always use several sources of information. Some people write to lead you astray. The online world exposed some interesting incidents that came out of the former Soviet Union before the attempted coup in 1991. Desinformation hurts everybody and comes from all sides. Even professional news agencies, like Associated Press, Reuters and Agence France-Presse, regularly stumble. Most news is written by journalists reporting what they have seen, read or heard. Their interpretation of the situation may be wrong. Supplement online news with what knowledgeable people say (by email or in conferences), when knowing the facts is important. Another point: Errors will occasionally be discovered and reported by the news sources, but always after the fact. Always store these reports in your archives, and make it a rule to search to the end when looking for something. Otherwise, you may never discover these corrections.
Calculating costs————————-Those living in Norway may read up to twenty-six pages of news fromAssociated Press in the United States and Financial Times (England)for US$ 0.64, or less.The trick is to dial long distance to a 9600 bps node in Swedenwhen the telephone company and CompuServe's non-prime time ratesare in effect.At 9600 bps, you may transfer text at up to 960 characters persecond. One page of text (size A-4) holds around 2200 characters.A typical news story is one to two pages of text.
| Users watching the 'taximeter' can use online services at a || very low cost. For many, global communication is almost free.|
Reading exactly the same news through another network or service, may cost you 300 percent more. Through yet another online service, the cost may double again. A full issue of the NewsBytes newsletter is around 150,000 characters, or 68 pages of text. Retrieving it from a local BBS typically costs me around 29 cents. Retrieving the full text from CompuServe would cost me over 500 percent more. Using NewsNet for the job, at 2400 bps through Datapak, would increase my current cost by more than US$30.00. The time of day may be important. Some services have different rates for access during the day, the evening, and the weekend. Use your calculator often.
When you pay by the minute ————————————— When using bulletin boards, phone charges are often the only cost items. Some boards require a subscription fee for full access to the system. Still, it is easy to calculate the costs of your calls. Divide the subscription fee by an estimated number of calls, and add to the cost of using the phone. The same applies to users of CompuServe. Their total cost is simply the sum of all connect charges, any network charges (to CompuServe and others), part of the basic subscription fee, and local phone rates (for direct dialing to the service, or to reach the network's node). Where a service uses a monthly subscription rate, add part of this to the time charges. Distribute the rate using an estimated number of online hours per month.
Example:You pay US$30/hour to access a service during prime time. Yourmodem speed is 240 cps.
Theoretically, if the data flows without pauses at system prompts, you can transfer 392 pages of text in one hour. Even when you deduct some characters due to stops in the transfer, the resulting transferred volume remains respectable.
To transfer one page of text takes around nine seconds (2200 characters divided by the speed, which is 2400 bps, or about 240 characters per second). The cost is nine cents.
A given binary file (a program) is 23552 bytes large. Using the XMODEM protocol, you can transfer it in about four minutes and thirteen seconds. The cost is US$2.10. To find the cost when paying by the minute is simple. Just calculate the cost per minute or second, and multiply by the estimated connect time.
On many services, it will take a minute or two before you can start to receive text or files. Disconnecting also takes a few seconds. Add this to the connect time when calculating costs.
Pauses and delays in the transfer can be caused by you or others, and may have a dramatic impact. It is particularly important to take this into account when comparing alternatives using different networks.
Example: Transfers to TWICS via Datapak at 9600 bps rarely gave me higher effective speeds than 100 cps. The reason was that the connection between the Japanese telcom network and TWICS went through a 1200 bps gateway.
A high speed connection to your data transporter's network does not guarantee a high speed connection to the remote computer.
I used to go through Datapak at 9600 bps to a computer center in Oslo. There, I was connected through a local area network to the host computer. The effective speed was rarely higher than 4800 bps. Calling direct gave twice the speed.
Try to measure the effective transfer speed before selecting a routing for your data. Transfer the same amount of text through various networks. If future transfers are likely to take place at a given time of day, test at that time. If your planned application is retrieval of programs, retrieve programs. If you want to read news, then read news from the services that you want to compare. When a network service charging for volume (like Datapak) will also be part of a comparison, measuring volume is particularly important. Do not assume that you know the answer in advance.
| NOTE: Always calculate the cost based on a fixed volume, like | | for the transfer of 1000 characters. This is particularly | | important when you need to use different transfer speeds to | | access competing services. |
Network load varies considerably throughout the day depending on the number of simultaneous users, and their applications. This also applies to online services. The load is normally lowest, when the bulk of the users are asleep, and during weekends. When the load is low, you get more done per minute.
Planning and self-discipline pays off ——————————————————- The actual cost of using a given set of services depends a lot on your self-discipline, the tools you use, and on how well prepared you are:
* If accessing manually, use "quick" commands rather than menus to move at maximum speed to desired sources of information.
* Do not set your services to be used with colors, sound, orspecial methods for displaying graphics, unless you have nochoice, or are willing to pay the extra cost. They increasethe volume of transferred text, and lower effective speed.
* Get the information and disconnect. It is expensive - andusually unnecessary - to read captured text while online.Log off to read. Call back for more to read, disconnect, andthen call back again.
* Learn how to write your mail offline, and send the letters"in a batch" to your mailbox. Your messages will often havefewer typing errors, be better thought out, and the costwill be considerably lower.
* Consider automating your communication (see Chapter 16). I use Bergen By Byte this way. A while ago, it gave me the following progress report: "Time on: 17 hrs 43 min, today 0 hrs 0 min, total 827 times." In average, I spend around 1.3 minutes per call. Yesterday, I was connected for 2:48 minutes. The result was 106 kilobytes' worth of conference mail.
Modem speed and cost —————————— 2400 bps is a sensible modem speed for some applications, and used to be a good starting point for new onliners. The benefits of using a faster modem may be marginal under the following conditions:
* When navigating the online service considerably reduces theeffective speed, and you access the service manually.* When you pay considerably more for access at higher speed.(CompuServe charges extra for 9600 bps access, but not much.)* When your networks do not offer higher speeds.* When the relative price of a faster modem in your countryis prohibitive.
On the other hand, a modem doing 9600 bps or more, does give you considerably faster communication. If doing things faster is more important than keeping costs down, then it is a wise investment. This is the case for me. Besides, often it is definitely cheaper. Your applications have a considerable impact on your costs. If you mainly use your modem for retrieval of programs and large data files from bulletin boards - and don't have to pay extra for volume - then higher modem speeds will immediately give reduced costs. A slower speed modem may also stop you from getting what you want. For example, there are several shareware programs on my board that users of 2400 bps modems are unable to download within their allotted 30 minutes per day.
When you pay for volume ———————————- Some network services, like Datapak in Norway, have high rates for volume, and very low rates for connect time. When using such services, automatic communication becomes less useful. Rather than connecting, getting a piece of information, disconnecting, and then going back for more, you may find it cost efficient to review menus and results while online. When paying for volume, the online service's menus become luxury items. Using quick commands for navigating is cheaper. Your comparisons will never be accurate when comparing with services charging for connect time. It is particularly difficult when the measure of volume is 'packets' rather than 'number of characters transferred'. Datapak and many other PDN services reports your sessions like this:
These numbers say that you have been connected to a service for 14 minutes and 55 seconds, that 537 data 'packets' have been received, and that 75 have been sent. Use these figures to calculate the cost of the call.
| One data 'packet' or segment contains up to 64 characters. | | Think of it as a measure of the number of lines. Each line can | | have a maximum of 64 characters. If you send the character A | | and a carriage return, then this also counts as a segment. | | | | Consequently, it is hard to use the Datapak record to estimate | | the real number of characters transferred. All we know is that | | 537 + 75 segments were transferred, and that 612 segments may | | contain up to 39,168 characters. |
When calculating the cost of a direct call, just the number of minutes counts. Use the time reported by the online service, and not your stop watch. CompuServe gives this type of report:
Thank you for using CompuServe!
Off at 10:11 EST 24-Nov-92Connect time = 0:15
If the size of your log file was 15 KB after the first test, and 11KB after the second, then just adjust the latter to compare (ActualCost/11*15). It is easy to compare services that only charge bythe minute.
More practical hints —————————— It is more expensive to call a service daily "to check the news," than to call it once per week to retrieve the same stories. Navigating by menus is more expensive than going directly to a source, or going there by stacking commands (i.e., combining quick commands into one). Many services let you read selective items in conferences by entering a search string. On my BBS, the following command
r extended 100+ c
lets you read all messages containing the search string 'extended' in the text starting with message number 100. If you forget the "c" parameter, the flow will stop after each message. This will reduce the average effective speed. Always use "nonstop" commands when reading stories, conference items, and other texts. Now, read the next chapter. Chapter 16: Automatic communication ===================================
Automatic data communication as a development strategy.
To get a lead on your competitors.To avoid duplication of effort.To reduce costs.To reduce boring and repetitive work.To avoid having to remember technical details.
Automatic communication is both for professionals and amateurs. First, because it keeps the costs down. Second, because it lets you do the job faster and safer.
We all have different needs —————————————- Automation will never be the same for everybody. Our needs are too different. Some get excited when a program can dial a bulletin board, retrieve a program, and then disconnect without them having to touch the keyboard. Some want an "answering machine" that can respond to and forward email when he or she is away from the office. Others want a communications system that can tap selected news sources, search databases, and do postprocessing on the retrieved material. For most professionals, doing things manually takes too much time. Time is better spent reading, digesting, and using, rather than on stupid technical retrieval work. Computers can do that. To others again, automation is a question of being able to use the online resource at all. If it takes 60 seconds to get a piece of information, it may be possible to get before running for the next meeting. If it takes 15 minutes, however, there may not be enough time. If you also need to read a help text to find out how to do it, you may not even consider it. The mind is full of other things right now.
| When using a system for automatic communication, you do not | | have to learn and remember online commands. The system will | | do it for you. |
The minimum solution —————————— Automatic data communication in its simplest form entails the following:
* One keypress to get the communications program to dial anumber, and send user name/password when the online servicerequests this information.
* Macro commands (like in a word processor) for navigatingthrough an online service, searching, and to send complexcommands by pressing one key.
Most communication programs have a macro language or a script language. You will probably never regret time spent on learning how to use these features. At a minimum, you should be able to have your system log on to a service automatically. Autologon spares you the task of remembering your user name and password. Besides, most people are only able to use the keyboard at a low speed. They easily get frustrated by having to correct typing errors.
Auto-logon with Procomm ———————————- Procomm is one of the most popular communications program in use today (see appendix 2). A Procomm script file is a text file, which can contain a list of commands for dialing and navigating on an online service. When writing a Procomm script for auto-logon, your first step is to list the commands that you believe required. Enter them in a text file (as DOS or ASCII text). In such scripts, you can test for the occurrence of a small piece of information that the online service is supposed to send at a given time (like the question "Password?"). When this information is found, Procomm can be set to send the proper response or command (here, your secret password). Scripts can be tied to your favorite online services through Procomm's dialing directory. Press a key to start the appropriate script file for access to a service. The following is a simple PROCOMM script file. It can be used to access my bulletin board in Norway. It assumes that your name is Jens Mikkelsen, and that the secret password is FOXCROOK4. You'll have to change this before testing.
;;Script file for auto-logon to SHS with PROCOMM and PROCOMM PLUS;WAITFOR "our FIRST Name? "PAUSE 1TRANSMIT "Jens^M"WAITFOR "our LAST Name? "PAUSE 1TRANSMIT "Mikkelsen^M"WAITFOR "ots will echo)? "PAUSE 1TRANSMIT "foxcrook4^M"WAITFOR "^JMore (Y),N,NS? "PAUSE 1TRANSMIT "n^M"WAITFOR "^JMore (Y),N,NS? "PAUSE 1TRANSMIT "n^M"WAITFOR "R] to Continue? "PAUSE 1TRANSMIT "^M"
It is not difficult. You probably understand a lot already. Here is the explanation:
* the ";" character at the beginning of a line identifies it as a comment line. Procomm is to ignore it. We use such lines for notes.
* WAITFOR "our FIRST Name? " has Procomm wait for the text string "our FIRST NAME?" from my BBS. It is a part of the question "What is your first name?".
* PAUSE 1 halts the execution of the script file for one second.
* TRANSMIT "Jens^M" sends the name "Jens" followed by a Return (the code ^M in Procomm).
* WAITFOR "our LAST Name? " makes Procomm wait for the question "What is your LAST Name?"
The script continues like this. In WAITFOR commands, we use part of the text that is displayed on our screen once the scrolling stops. Make sure that the search term is unique. It must not appear elsewhere in the text coming from the host computer. If it does, your name and password may be sent too early. You can call the script HORROR.CMD, and attach it to the entry for my board in your Procomm phone directory. When you call it the next time, Procomm will execute the commands in the file and "turn the keyboard over to you" when done.
Macros in Procomm ————————- Above, we used a script to log on automatically to a service. When Procomm gives us access to the keyboard again, we must continue manually. What we want to do online varies. Sometimes, we want to read new messages in conferences. In other cases, the purpose is to check new programs in the file library. If we find programs of interest, we may want to download them. Shorthand macros can help you do this faster and safer. For example, one macro can take you quickly to a conference for new messages. You can make Procomm start this macro whenever you press ALT-0 (keep the ALT key down, then press 0). You can have the macro key ALT-1 send other commands when in the file archives. When I started using MS-DOS computers for data communications, PC-TALK became my favorite program. It has many of the same macro capabilities that Procomm has. With PC-TALK, I did autologon to NewsNet. Macro number one sent commands that gave me the contents of various newsletters. Macro #2 picked up the contents in another group. Macro #3 picked up stories from my mailbox, and macro #4 logged me off the service. My mission was completed by pressing four or five keys.
Automating the full task ———————————— It's a long way from automated logon scripts and the use of macros to automating the whole task. The major difference is that with full automation, you do not have to look at the screen while the script is working. You can do other things. Sometimes, you may not even be present when the job is being done. On a typical morning, I go directly from bed to my office to switch my communications computer on. While I visit the bathroom, my communications program calls three online services, retrieve and send information. When the script has disconnected from the first service, which is my bulletin board, it analyzes the received data. I want an alphabetic list of visitors since my last visit, a sorted list of downloaded programs, and names of those calling in at 9600 bps or higher. Sometimes, the unexpected happen. There may be noise on the line, or a sudden disconnect. Usually, my script can solve this without manual intervention. It is therefore allowed to work unattended most of the time. When I get to my office after breakfast, it is all done. My communications program is set for reading and responding to today's email. I can sit down, and immediately get to work. After having written all my replies, I say "send" to my system. For me, it's time for another cup of coffee. I am not needed by the keyboard while my mail is being sent. This is what an automatic communications system can do. My scripts also help plan and prepare online visits, and ease my work by postprocessing results.
| When your communication is fully automated, you need not | | read incoming data while it scrolls over your screen, and | | then again after logging off the service. You do it only | | once. |
How to get it? Here are some alternatives:
Alternative 1: Write your own system —————————————————— You can write procedures for powerful script-driven programs like ProYam (from Omen Technology) and Crosstalk MK IV. I started writing scripts for ProYam over seven years ago. The system is constantly expanded to include new services, refined to include more functions, and enhanced to become more robust. The scripts make my system work like an autopilot. It calls online services, navigates, retrieves and sends data. Postprocessing includes automatic reformatting of retrieved data, transfers to various internal databases, statistics, usage logs, and calculation of transfer costs. Such scripts can do quite complex operations online. For example, it can
- Buy and sell stock when today's quotes are over/undergiven limits,- Select news stories and other types of information basedon information found in menus or titles.
Script writing is not for everybody. It is complicated, and takes a lot of time. Therefore, it is only for the specially interested. On the other hand, those going for it seldom regret. Tailor- made communication scripts give a wonderful flexibility. The software does not cost much, but again, it takes a lot of time!
| Do not use large and complex script files before you know the | | online service well. The scripts let you do things quicker and | | safer, but there is always a possibility for unexpected | | problems. | | | | Test your scripts for a long time to make them robust by | | "training" them to handle the unexpected. Leave them to work | | unattended when you are reasonably certain that they can do | | the job. - It may take months to get to that point. | | | | Build a timeout feature into your scripts, so that they don't | | just hang there waiting for you after an encounter with fate. |
Alternative 2: Use scripts made by others ————————————————————- Some script authors generously let others use their creations. Earle Robinson of CompuServe's IBM Europe Forum, share his ProYam scripts for automatic usage of CompuServe with others. They are available from the IBM Communication Forum library. Enter GO XTALK on CompuServe to find advanced script files for Crosstalk Mk.4. ZCOMM and ProYam scripts for visiting my board automatically can be freely downloaded there. They split access up into these three phases:
Phase 1: Menu driven offline preparation.Phase 2: Automatic logon, navigation through the system, andautomatic disconnection.Phase 3: Automatic offline postprocessing.
You will find scripts for other programs on many online services.
Alternative 3: Special software ———————————————- Several online services sell communication programs with built-in functions that provides you with automation. They can have offline functions for reading and responding to mail. The degree of automation varies. There are also many programs written by third parties. Most programs assume that you use 'expert' as your default operating mode on the online service. TapCIS, Autosig (ATO), OzCIS, CISOP, CompuServe Navigator (for Macintosh), AutoPilot (for Amiga), ARCTIC (for Acorn Archimedes), and QuickCIS (for Atari) are popular choices on CompuServe. TapCIS is my personal favorite. (CIM does not offer much automation!) Aladdin is for GEnie. It automates your use of RoundTables (conferences), file areas, and mail. Dialog users turn to Dialog- Link. Nexis News Plus (for Nexis, US$50) has pull-down menus and detailed selection of commands. This MS-DOS program helps users set up detailed search commands before logging on to the Mead Data Central. Your search results will be downloaded automatically. Personal Bibliographics Software, Inc. (Ann Arbor, Mich, U.S.A. Tel.: +1-313-996-1580) sells Pro-Search to Dialog and BRS users (for Macintosh and MS-DOS). Pro-Search will lead you through menus to find information on both services. It translates your plain English search commands into the cryptic search language used by the services. It logs on automatically, connects to these services, finds your information, and shows you the hits.
Alternative 4: Offline readers ——————————————— The alternatives above have one important weakness. Noise on the line can prevent the "robot" from doing the job. All it takes is for noise to give a prompt another content than is expected by your program or script (as in "En@er a number:" instead of "Enter a number:"). You can avoid noise problems by using get commands (see Chapter 15), and by making the online service use its minimum prompts ('expert mode') . Still, this does not give full protection. The best is to let the online service do the navigation. Think of it as logging on to run a batch file on the remote computer. Combine this with automatic transfers of your commands, transmitted in of one stream of data with automatic error correction (in the software and in the modem), and you have a very robust system. The program logs on to the service. Then the service takes over. It registers your user identity, checks your user profile for personal interests, retrieves and packs all messages, news and files into one compressed file, and sends it to you at high speed. Your outgoing messages, search commands, commands to join or leave conferences, and more, are transferred to the remote computer in a similar packet (compressed file). When received by the remote computer, it unpacks the transfer file and distributes messages and commands to various services following your instructions. Your "physical" contact with the service is when your modem is disconnected. The help menus that you read belong to your program, and not the online service. You read and respond to mail in a reading module (ref. the term "offline reader"). Some offline readers give the caller access to more tools than is available on the online service itself. They may have spelling checkers, multimedia support, let you use your favorite editor or word processor, and offer various storage, search, and printing options. Using offline readers is probably the easiest, cheapest, and safest way of using online services. These "readers" are popular among bulletin board users, and some commercial services are also starting to accommodate them. There are many offline reader programs. The most advanced take over completely upon logon, and manage transfers of commands and compressed information files to and from the host. (Example: Binkley Term on FidoNet) Global Link is an offline reader for EcoNet. Bergen By Byte offers the BBS/CS Mail Grabber/Reader, a script system used with the communications program Telix and the service's "auto-get" function. The most popular systems on the PCBoard based Thunderball Cave BBS are Offline Express, Megareader, Session Manager, Rose Reader and EZReader. These are used with scripts written for various communication programs. Some of them have built in communications (and script) modules. EZReader from Thumper Technologies (P.O. Box 471346, Tulsa, OK 74147-1346, U.S.A.) lets users retrieve mail from several online systems using transfer formats such as QWK, PCBoard capture files, ProDoor ZIPM files, XRS, MCI Mail, and others. Cost: US$49 (1992). 1stReader from Sparkware (Post Office Box 386, Hendersonville, Tennessee 37077, U.S.A.) is my personal favorite for accessing Qmail based online systems.
| Note: Some offline readers contain all the features required | | for fully automated communications. Some bulletin boards allow | | up- and downloading to start right after CONNECT. | | Off-Line Xpress, an offline mail reader for QWK (Qwikmail) | | packets, does not contain a communications module. It just does | | pre- and postprocessing of mail packets. | | You can use the Off-Line Xpress as one element in a larger | | automated system. For example, a system for access to PCBoard | | bulletin boards may consist of Off-Line Xpress software, PKZIP | | and PKUNZIP (popular shareware programs to compress/decompress | | mail packets), the QMODEM communications program, and a script | | to navigate to/from the QWK packet send and receive area on the | | BBS. | | 1stReader (version 1.11) contains a powerful script based | | communications module. It lets you compose replies, set search | | commands, subscriptions to services, add and drop conferences, | | and enter download commands offline. |
Automatic automation —————————— We have explained how to write scripts with Procomm. However, there are simpler and quicker ways. Many communication programs can make scripts automatically using a learning function. It goes like this: Start the learning function before calling the online service. Then log on, navigate to the desired services, do what you want to automate, and disconnect. The learning feature analyzes the received data and builds a script file for automatic communication. If you call again with the new script, it will "drive the same route one more time." ZCOMM and ProYam have a learning feature. This is how I made a script for accessing Semaforum BBS using ZCOMM:
ZCOMM asked for a phone number. I entered +47-370-11710. It asked for speed, and I entered 2400 bps. Next, I had to choose one of the following:
(1) System uses IBM PC (ANSI) line drawing(2) 7 bits even parity(3) 8 bits no parity
My choice was 1.ZCOMM dialed the number. When the connection was established,I entered my name and password, navigated to the messagesection, read new messages, browsed new files in the library,and entered G for Goodbye. This was the "tour" that I wanted toautomate.When disconnected, I pressed the F1 key. This prompted thelearning process based on a record of the online tour. The logdescribed everything that had happened in detail, including mypauses to think. Now I was prompted by the following question:
'newscr.t' exists. Append/replace/quit?
I selected append. Then:
Do you want this script file as a new entry in yourtelephone directory (y/n)?
I entered "y," and named it "semaforum." After a few seconds,my new script was ready:
Your new script is in the file 'newscr.t' !!You can append the file to your current script file(for example PHODIR.T) or have the commands executed byentering:call semaforum.newscr.t
It was time to test the new wonder. I entered
call semaforum.newscr.t
at the ZCOMM command line, hit the Enter key, and off it went. ZCOMM called the BBS and repeated everything - at far higher speed than I had done it manually. It went on-hook as planned when done.
Limitations —————- Auto-learn programs can create a script file that let you "drive the same route." For some applications this is enough. For others, it's just part of the way. You have to refine the script manually to get what you want.
Example:If you call my bulletin board with an auto-learned script madeyesterday, chances are that everything works well. If you calltwice on the same day, however, you're in for a surprise. Theboard greets you differently on your second visit. You will notget the menu of available bulletins. It will take you directlyto the system's main menu. Your script must take this intoaccount.
On most online services, many things can happen at each "junction of your road." At one point in one of my scripts, up to twenty things may happen. Each event needs its own "routing." Twenty possible events are an extreme, but three to four possibilities at each system prompt is not unusual. All of them need to be handled by your script, if you want it to visit online services unattended while asleep. It is quicker and simpler to use other people's scripts and programs, although this might force you to use a different program for each service. Personally, I prefer offline readers on services where such are able to do the job. On other services, I usually depend on my own tailor-made scripts. Chapter 17: Gazing into the future ==================================
Thoughts about things to come.
Newspaper of the future —————————————- Some years ago, Nicholas Negroponte of Massachusetts Institute of Technology, said that today's newspapers are old-fashioned and soon to be replaced by electronic "ultra personal" newspapers. "If the purpose is to sell news," he said, then it must be completely wrong to sell newspapers. Personally, I think that it is a dreadful way of receiving the news." MIT's Media Laboratory had developed a new type of electronic newspaper. Daily, it delivered personalized news to each researcher. The newspaper was "written" by a computer that searched through the news services' cables and other news sources according to each person's interest profile. The system could present the stories on paper or on screen. It could convert them to speech, so that the "reader" could listen to the news in the car or the shower. In a tailor-made electronic newspaper, personal news makes big headlines. If you are off for San Francisco tomorrow, the weather forecasts for this city is front page news. Email from your son will also get a prominent place. "What counts in my newspaper is what I consider newsworthy," said Negroponte. He claimed that the personal newspaper is a way of getting a grip on the information explosion. "We cannot do it the old way anymore. We need other agents that can do prereading for us. In this case, the computer happens to be our agent."
The technology is already here. Anyone can design similar papers using powerful communication programs with extensive script features. I have tried. My test edition of The Saltrod Daily News did not convert news to sound. It did not look like a newspaper page on my screen. Not because it was impossible. I simply did not feel these 'extras' worth the effort. My personal interest profile was taken care of by my scripts. If I wanted news, the "news processor" went to work and "printed" a new edition. On Tuesdays, Thursdays, and Saturdays, I got an "extended edition." This is a section from my first edition:
"Front page," Thursday, November 21.Under the headline News From Tokyo, the following items:
These articles were captured from Kyoto News Service throughDown Jones/News Retrieval.
The column with news from the United States had stories fromNEWSBYTES newsletters:
* DAY ONE COMDEX.* IBM'S PRE ANNOUNCEMENT OF "CLAMSHELL"* AT&T TO JUMP IN SOONER WITH LAPTOP COMPUTER* COMMODORE THIRD CONSECUTIVE QUARTERLY LOSS* 2 ZENITH UNVEILS TOUCH-SCREEN* HP's EARNINGS DROP
Hot News From England came from several sources, includingFinancial Times, and Reuters (in CompuServe's UK News).Headlines read:
"Page 2" was dedicated to technology intelligence. "Page 3" had stories about telecommunications, mainly collected from NewsNet's newsletters. "Page 4" had stories about personal computer applications.
As the cost of communicating and using online services continues to decrease, many people will be able to do the same. This is where we are heading. Some people say it is too difficult to read news on a computer screen. Maybe so, but pay attention to what is happening in notebook computers. This paragraph was written on a small PC by the fireplace in my living room. The computer is hardly any larger or heavier than a book. (Sources for monitoring notebook trends: NEWSBYTES' IBM and Apple reports, CompuServe's Online Today, and IBM Hardware Forum.)