lauren@vortex.UUCP (Lauren Weinstein) (08/16/85)
I disagree with this concept. It's IMPOSSIBLE to keep followups from being posted to such groups, and I frankly am beginning to doubt that the network can handle much more stuff being sent to every site when it really only needs to go to a relatively few individuals. A more reasonable scenario is to post ideas for projects in the groups that deal with those topics, then collect the names of interested parties and run a mailing list after that. If you want to do something with a C compiler (free or commercial) then the idea can go to the C newsgroup, and the author can collect the names of interested parties and proceed from that point with a specific mailing list. If you have an analog project, use the existing analog group. An advantage of using already existing groups is that they (by definition) already reach the people interested in those topics. This seems far superior to creating another group that will inevitably generate random queries and followups and the usual amount of newsgroup "noise." There are group projects going on now via mailing lists, without forcing the entire network to pass around all that traffic. If the volume of such proposed projects became extremely immense, then a separate newsgroup *might* make sense--but even then I would tend to say that posting such queries to the specific technical group in question would be superior. Network traffic is starting to go suddenly out of control again. 100K+ postings to net.tv.drwho. 100K postings (old bibliographies) to net.research. Endless, useless muck burying an occasional gem in net.bizarre. Intense use of followups of included text in both technical and non-technical groups. One local site found one of their dialups lines virtually CONTINUOUSLY in use trying to handle netnews feeds and is on the verge of turning off netnews completely. One of the major problems with creating new groups is that each group carries with it a certain amount of new "noise"--users who don't know what they are doing, inappropriate queries, voluminous back postings and included text, etc. A new group tends to be an excuse for all sorts of garbage--look again at net.bizarre, which started out well enough but has quickly become bogged down in every bit of computer trash that people have laying around. In a short time it will probably be just like net.flame--99% useless--and we *really* didn't need to double the amount of net-flame-like traffic in an environment where major sites are bogged down, other sites are cutting off some or all groups, and we're STILL faced with enormous growth in terms of numbers of submitting sites. So, to the extent that it is possible, let's TRY to use existing groups instead of creating new ones. An occasional project idea for, for example, a C processor can go to the C groups, just as other sorts of queries could go there and reach people interested in C. We don't create a general-purpose "query" group to which all queries, on whatever topics, should be posted; we instead post our queries to the specific group of interest, I feel strongly that the same principle applies here. It would apply even if the network had infinite bandwidth and cost nothing whatever to run, which it most certainly does not. --Lauren-- P.S. Sorry for the lateness of this reply, but I haven't received much netnews for three days due, you guessed it, to a clog at a major site. --LW--
umdhep@eneevax.UUCP (Todd Aven) (08/18/85)
I think that all the important objections to net.software.projects about cluttering the lines with junk can be blown away by having the group moderated, as I proposed before. The net may be cluttered, but if it can carry an un- moderated trash group like net.bizarre or net.flame, then it can handle a moderated group like mod.projects to actually perform a service to a number of users. Networking is a great means to provide cooperation in projects indepen- dent of physical distance, and this group would enhance the capabilities of the network in the direction originally established. I think that we can bear the weight! ============================================================ |Todd Aven MANAGER@UMDHEP.BITNET | |Softwear Sweatshop AVEN@UMCINCOM (arpanet, bitnet)| |High Energy Physics UMDHEP@ENEEVAX.UUCP | |University of Maryland | |College Park, MD 20742 (301)454-3508 | ============================================================
lauren@vortex.UUCP (Lauren Weinstein) (08/19/85)
The fundamental problem is that the project-type messages we're talking about don't belong in their own groups -- they belong in the specific group for the project topic of interest. C projects in net.lang.c, analog projects in net.analog, etc. This makes sure that the people most interested in the topic, by definition, see the messages. Setting up separate group(s) would be almost literally like setting up a net.query group for people to send all questions on any topic to, rather than having them send their questions to the specific topic-oriented (net.lang.c, net.analog) group of interest. --Lauren--
umdhep@eneevax.UUCP (Todd Aven) (08/21/85)
The error in Lauren's reasoning (projects should be discussed in the topical newsgroup) is that projects don't necessarily start with a chosen language. The idea usually develops in a vague manner, and after some discussion, a medium in which to develop the idea is chosen. This is exactly why mod.projects is necessary. Andrew, either create it or I will. (No I don't know how, but I want it and I'll find out how). Todd Aven Softwear Sweatshop High Energy Physics
peter@baylor.UUCP (Peter da Silva) (08/22/85)
> The fundamental problem is that the project-type messages we're talking > about don't belong in their own groups -- they belong in the specific > group for the project topic of interest. C projects in net.lang.c, > analog projects in net.analog, etc. This makes sure that the people > most interested in the topic, by definition, see the messages. > > --Lauren-- OK, everyone... how about some projects? -- Peter (Made in Australia) da Silva UUCP: ...!shell!neuro1!{hyd-ptd,baylor,datafac}!peter MCI: PDASILVA; CIS: 70216,1076
lauren@vortex.UUCP (Lauren Weinstein) (08/23/85)
Gee, it's SO NICE to learn that Todd has become God of the Network. "Either I will create it or so and so will!" he says. That sort of attitude will KILL THIS NETWORK. Ya wanna see what's gonna happen when everybody and his brother decides that 20 or 30 people want to discuss some random topic and simply declare they're going to create the group? As a practical matter, the only person who really should be creating groups is the person who maintains the master group lists. "I want it so I'll create it," he says. Sounds like a little child who wants candy. Note that I'm not saying anything about the pluses or minuses regarding the groups at this point--I'm just pointing out that the attitude of "since I want it I'll create it" is infantile and dangerous for the structure of the entire network. Not just in this particular case, but in general as other people decide they can create whatever groups THEY want based on their own nebulous criteria. I will be generating a long message discussing the reasons why the projects group is a bad idea, point by point, shortly. In the meantime, I have received something like 30 messages from people saying they think the group is a bad idea--but many have expressed the fear of saying this to the network publicly since they're afraid (possibly quite rightly) that they'll be harrassed. I encourage these people to say publicly what they've been saying to me in private mail. --Lauren--
jaw@ames.UUCP (James A. Woods) (08/24/85)
# "no one can make you feel inferior without your consent" -- eleanor roosevelt again, lauren, in his very admirable (but nevertheless quixotic) role as "the voice of sanity" on usenet, has perceived the essence of the matter. net.projects is simply too general, and would inevitably produce the flotsam now seen even in the likes of (the presumably staid) net.database. mod.projects is more like it, but even here, mailing lists arising from individual technical groups would do the job fine. for example, last year, folks talked about news compression a bit in net.news. someone put the question to net.news about whether anyone was working on same. the ieee computer article came out, code was posted to net.sources, and an international project to improve the utility was carried out -- first, by a small cadre of correspondents connected via a complete graph, and then, through the traditional moderated "reflector" style mailing list. if the "compress" discussions were to be net-wide, then thousands would be bored by implementation details. (to polish an apple here, the intrepid lempel-ziv freaks will always be in the unique position of having a "negative" byte-count associated with their net verbiage.) i myself am not sure whether the freeway hypothesis holds (more groups cause more congestion), or whether traffic is a linear (or asymptotic) function of uucp logins independent of group cardinality. i'm only protesting about net.projects here, not new group formation. -- james a. woods (ames!jaw)
peter@baylor.UUCP (Peter da Silva) (08/29/85)
> Gee, it's SO NICE to learn that Todd has become God of the Network. > "Either I will create it or so and so will!" he says. That sort > of attitude will KILL THIS NETWORK... Right on! I disagree with what you say about net.software.projects, but I'd rather see it out in the open where I could debate (hopefully without flaming) it. -- Peter (Made in Australia) da Silva UUCP: ...!shell!neuro1!{hyd-ptd,baylor,datafac}!peter MCI: PDASILVA; CIS: 70216,1076