bobvan (11/24/82)
A significant percentage of the articles I read on the net request information on a particular topic. Those who post these articles frequently become a sort of "instant expert" in the field. If lucky, the net gets a survey article in return. Has anyone every considered a newsgroup that would carry a regular listing of topics of current interest and the net address of the last person to conduct a survey of the topic? The idea would be that you look at the current list BEFORE posting a query to net.wanted. If you find an expert in your area of interest or a closely related one, a piece of mail can probably bring a survey, without bothering the whole net. I guess I'm getting tired of seeing the n'th request for information on a given topic. I know that netnews has produced "experts" on the following topics: - uucp bugs - laser printer vendors - unix spread sheets - S-100 floppy disk controllers - C compilers for the foo micro - UNIX on the foo micro - mice vendors - track ball vendors - available UNIX device drivers This list was produced from recent memory and is far from exhaustive. I doubt if we can attach any sort of "obligation" to the act of posting a request for information, though it is an interesting thought. The maintenance of the current "expert list" and its regular posting would require a little of someone's time. This person would become the "expert expert." The obvious starting point for a group name discussion is net.experts. I consider this a "raw" suggestion that still needs work before I'm happy with it. Much of the value in this information is that it is current. The "high-tech" information we are dealing with here tends to age at an incredible rate -- the market surveys in particular. The success of such a group would certainly hinge upon having a good "expert expert" and upon acceptance and correct use by net users. What I am proposing would make a significant change in the way the net is used to disseminate information and it would take some time for the manuals and the never-expiring articles to be updated. Another problem is that net addresses can tend to be rather dynamic. My own address will change about the first of the year and that is one of the reasons I can't volunteer for the job of "expert expert". The expert list will have to be updated as things change. Feedback from frustrated users of the list would be a great help here. Perhaps a good way to gauge the potential for success of the proposed group would be to solicit mail responses from experts willing to have their names on such a list. I will be happy to collect such mail and build the initial list, but I cannot serve as the "expert expert" for the reason mentioned above. If you wouldn't mind answering an occasional survey request, please send me your name, net address, and area of expertise. I use the word "expertise" here rather loosely. If you've spent more than a day researching something, you can probably save another netnews reader at least that much time by sharing what you've found. A similar request for experts should go to net.general and to net.unix-wizards (for the ARPA gateway) once a more permanent "expert expert" has been found (any volunteers?). Please contribute your thoughts on this idea. Send mail or post as appropriate (I think it safe to trust the judgement of net.news.group readers). Thanks for your input, Bob Van Valzah (...!decvax!ittvax!tpdcvax!bobvan)
sjb (11/24/82)
This idea, while good in the fact that it would cut down net traffic, doesn't sit well with me for the reason that you'd be binding people to just a few opinions on one subject rather than thousands. I can more easily make a decision on something I know nothing about when a few thousand people give me good advice on it rather than when one or two people give me advice. The net is here to allow people to communicate openly with one another and for discussing, asking, and answering these questions. Why reduce it to a place you go to just to find out who to ask for something?
bobvan (11/27/82)
A recent article by alice!sjb criticizes my net.experts proposal. Comments in this article lead me to believe that I haven't properly explained my proposal, so some further clarification is in order. From sjb's article: This idea, while good in the fact that it would cut down net traffic, doesn't sit well with me for the reason that you'd be binding people to just a few opinions on one subject rather than thousands. I envision net.experts as the starting point in any search for information -- not the ending point. I am NOT suggesting that people be forbidden from posting inquiries to the net, but rather that they check net.experts FIRST to see if someone else may have already done a survey. To be sure, many people will need information that is not covered in the current experts list. I'd like to see more surveys on the net -- not fewer. I'm just tired of seeing the fourteenth request for "the best home computer under $300" or "the best laser printer to work with troff". Now I'd like to be specific about what such a survey might contain. Surveys should consist mostly of facts, net addresses of those who may have further information, manufacturers address, and finally the surveyors opinions on the current "best" along with the reasons for his choice. The point here is that this net BUILDS experts. By my definition, an expert is not someone who sits in a cave and reads Electronics, professing his opinion on any subject of interest. Instead, an expert is made when one individual requests information, receives several replies from the net, digests and summarizes this information, extracts information from manufactures data sheets, and finally publishes the result with his conclusions and their justification. I've included a shortened copy of a recent summary I did covering the laser printer market. I asked for information on laser printers several weeks ago. I apologize for being lax in getting a summary posted, but better late than never. I am very grateful to those who responded to me, especially Rick Furuta. I'll give a brief summary of my findings here and will mail the raw information to anyone who is interested. There are three printers that we seriously considered: the Symbolics LGP-1, the Imagen IMPRINT-10, and the Xerox 2700. We did not consider the DEC LN-01 because it is not yet available. When it does arrive, I am told it will use the Xerox 2700 printing engine. The Imagen and Symbolics both use the Cannon printing engine and are roughly equivalent. Both come with software interfaces for troff and plenty of fonts. TeX and Scribe drivers either are or will be available. Both can be connected to your machine over a standard 9.6k or 19.2k serial line. Symbolics also has interfaces for a Centronics standard 8-bit parallel and high speed 16-bit parallel. Imagen has been shipping for over a year, as of two weeks ago Smybolics had yet to ship their first printer. Imagen costs ~30k$, Symbolics about 25k$. Imagen has service available in some cities, Smybolics is negotiating with a service contractor. Imagen will do 240X480 dots/inch, Symbolics 480/240. In summary, Symbolics seems to have a slightly better product for less money but it is not an established product. We have ordered one. The Xerox is in a different class. Cost is 19k$, lease is about $650/mo., resolution is 300X300, full service support is available, interface is through serial line only. It has several nice fonts and can draw lines with escape sequences, but you can't get at the individual dots on the page. You can define your own fonts and down load them. Naturally, there is no troff interface. It seems that nroff -Txerox might be more appropriate. Notice that it consists primarily of facts and objective comparison. Since I posted this to the net, I did not include manufacturers addresses or net addresses of contributors. I have forwarded this information by mail to those who asked for it. This summary was posted to net.unix-wizards less than a month ago and just yesterday, I saw yet another article there requesting information on laser printers that could be used with troff. I mailed the author of the article a copy of my summary because it seemed clear that he'd not seen it -- perhaps it got lost at the gateway (ARPA address). In any event, this is the sort of senseless sort of duplication of effort I'm trying to prevent. To summarize, I do not seek to limit you to the opinions of one person, but rather to have one person COLLECT and distribute the opinions and information from many people. If this information were inadequate, I encourage you to use the net to get the information you're after as long as you summarize your results. Bob Van Valzah (...!decvax!ittvax!tpdcvax!bobvan) P.s. I just saw an article that went something like "Does anyone know the net address of the guy who did the survey on UNIX work processing packages? I didn't need it at the time it went by, but now I do." This is just the sort of need that net.experts could fill. PP. S. Perhaps my error was in suggesting the term "expert". This term may place too much emphasis on opinions rather then collected facts. Maybe net.surveys would be a better name?
sjb (11/28/82)
I still don't see how net.experts (or whatever it would be called) will solve the problem of countless queries for information that has already been asked for or posted. The main problem is network delays, i.e. site x may not get an article from site y for days, therefore users on site x may post queries or followups to articles from site y that were already answered since they never saw the original answer first. net.experts won't solve this, as it will have just the same problem. As for things that were asked months or years ago, it still won't solve the problem since most sites get rid of news that's more than 2 weeks old and you'd be asking and- ably) to do.
bobvan (12/01/82)
My proposal for an "expert list" has now been rattling around the net and around inside my head for about a week now. I've received about a half dozen encouraging replies by mail and two discouraging replies on the net. Only one person volunteered their own expertise in a field and no one volunteered to be the "expert expert". As a result of this, I've come to the following conclusions: My original idea was much too grandiose. The idea does not warrant its own newsgroup. Re-writing manuals and posting never expiring articles is far too much work for it to be readily accepted. The term "expert" has too many negative connotations. Nobody wants the responsibility of being an expert. These conclusions leed me to the following new proposal: someone maintains a list of all surveys and survey requests posted to the net. The list is regularly posted to net.general, with expiration set to the next scheduled posting date. The list is called the "survey survey" and is distinctively titled so that it can be easily n'd. For each survey, the list contains: the date of the survey, the name of the surveyor, the net address of the surveyor, and the subject of the survey. In order to make the idea more concrete, I have grep'd thru some of the articles on line here and produced a crude survey survey. It follows the signature on this article. Be forewarned that none of the people listed have consented to have their names appear on the list. This would be an essential part of a continuing survey survey. Here are some of the problems that I see with the list as it stands now: The current format (four columns, tab separated) is easily machine readable, but inconvenient for humans. The division of the list into topics complicates this further. The list contains both USENET and ARPA addresses -- no mention of the "secret gateway" is made. Perhaps the list should be posted to net.unix-wizards so that ARPA people can both benefit and contribute. I am leaning toward a fresh posting every two weeks, at least until the list "matures." I welcome your comments on the refined proposal above. Thanks to all who contributed constructively to this proposal. Bob Van Valzah (...!decvax!ittvax!tpdcvax!bobvan) net.general net.general net.general net.general 15 Nov 82 Kirk Glerum uw-beaver!kirkg inexpensive "home" lasers 17 Nov 82 Hal Perkins cornell!hal music editing/printing software 26 Nov 82 Jay Sanders hssg40!jay word processing software for Unix 29 Nov 82 Chris Hibbert grkermit!chris solar home controllers 19 Nov 82 Jerry Schwarz eagle!jerry Usenet Etiquette 18 Nov 82 Curt Stephens tekmdp!curts purpose of each USENET newsgroup 20 Nov 82 Ian Darwin utcsstat!ian Spreadsheet Calculators on Unix 25 Nov 82 name unknown micomvax!george commercial relational database packages compared to INGRESS 29 Nov 82 Dave Ihnat ihuxx!ignatz How many people read netnews net.unix-wizar net.unix-wizar net.unix-wizar net.unix-wizar 29 Nov 82 John P. Nelson genradbolton!john documentation for 4.2 bsd IPC mechanism 16 Nov 82 Bill Lee lee@UTEXAS-11 VMS vs. Unix benchmarks 18 Nov 82 Richard Stevens kpno!stevens what are UNIX versions 4 & 5 ? 21 Nov 82 name unknown utcsrgv!peterr C bug logs 27 Nov 82 Charles F. Von Rospach CHUQUI@MIT-MC porting PDP-11 C to a VAX 24 Nov 82 Alan S. Watt ittvax!swatt available unix device drivers 29 Nov 82 Bob Van Valzah tpdcvax!bobvan inexpensive laser printers net.micro net.micro net.micro net.micro 28 Nov 82 Michael Wagner utcsstat!wagner Moving Intelligence into the Terminal 28 Oct 82 David Hinnant tucc!dfh UCSD Pascal internals 04 Nov 82 aron shtull-trauring presby!aron ADA(TM) for z80s 12 Nov 82 Guy Riddle pyuxmm!ggr how to make an intel 8274 work 16 Nov 82 John Reese ll1!jmr1 JRT Pascal (for CP/M) 12 Nov 82 Mike Knudsen ihnss!knudsen TRS Color Computer (comparison to others) 17 Nov 82 Bill Rizzi RIZZI@Usc-Isib file management (DBMS) packages written in C net.singles net.singles net.singles net.singles 27 Nov 82 Wm Leler unc!wm favorite places to meet persons of the appropriate sex 27 Nov 82 Wm Leler unc!wm experiences with "swinging" parties net.audio net.audio net.audio net.audio 29 Nov 82 Ira Idelson ariel!ili How to Pick A Turntable