Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×
Programming IT Technology

Overcoming the Network Effects? 30

Posted by Cliff from the this-is-a-job-for-*gasp*-marketing dept.
paul_harrison asks: "I am trying to introduce a new P2P protocol. It's technically superior in several respects to existing protocols, but there's one big problem: too few people using it. Now this is not a new problem, there's even a name for it, "Network Effect". It crops up all over the place: which websites become popular, which formats and protocols people use, which operating systems people use, even which side of the road people drive on... So my question is this: how do things like these overcome network effects and become popular?"
This discussion has been archived. No new comments can be posted.

Overcoming the Network Effects? More

Overcoming the Network Effects?

Comments Filter:

  • Obvious (Score:3, Funny)

    by Clue4All ( 580842 ) on Thursday May 30, 2002 @11:46AM (#3609668) Homepage
    Post a story on Slashdot about it. Too bad it didn't make it onto the front page, you'd be all set.
  • ...by word of mouth. Get your friends on it, your families, your co-workers, thier friends, etc. It will soon spread.

  • missing key features (Score:3, Funny)

    by tps12 ( 105590 ) on Thursday May 30, 2002 @11:48AM (#3609678) Homepage Journal
    You need to add random crashes, spyware, and poor interface design before it will catch on. These add to the thrill and danger of pirating music and software.
  • Advertising and abuse of monopoly power are the two main ways.
  • If I knew the answer to this, everyone would know who I am and worship me.

  • You contradict yourself (Score:5, Insightful)

    by whoda ( 569082 ) on Thursday May 30, 2002 @11:57AM (#3609753) Homepage
    You say: "Which means no censorship, no entry taxes, no one booting you off the network, and no weak point which can break the whole system."

    Then at the end you state: "In particular, if you try to share snuff or child pornography, I will be able to work out you IP and from that your location and identity. And I will report you to the police."

    So, which is it? P2P with no censorship, no booting, or P2P With censorship and booting?

    • No contradiction at all... (Score:4, Informative)

      by damiangerous ( 218679 ) <1ndt7174ekq80001@sneakemail.com> on Thursday May 30, 2002 @12:38PM (#3610042)
      No one has the ability to remove you from the network, or control ("censor") what you share. However, you are not completely anonymous on the network, and if you share something illict you can be found if someone really wants to (which this fellow apparently does).
      • I believe (I don't know whether it's possible to boot someone, maybe it's like a network command or something) that the point is to allow freedom to send stuff like music, videos, etc... but not allow sick things like child porn.
    • Oh no - he's censoring child pornographers. Heavens to betsy.

      Unless you're terribly broken up about some sick bastard being sent to jail for child porn, get over it.
      • Oh no - he's censoring child pornographers. Heavens to betsy.

        Unless you're terribly broken up about some sick bastard being sent to jail for child porn, get over it.

        No, we do not like child pornographers.
        No, I wouldn't be torn by any of them going to jail.
        Hell, I think they all should be shot.

        But, that isnt the point. The author sas there will be no banning. But he tells us he will ban child pornographers?
        Once he changes his wording, I will be happy, but I dunno if the rest of the people will be.

        • The author sas there will be no banning. But he tells us he will ban child pornographers?

          You misunderstand. There is no contradiction. You can share all the child porn and snuff all you want, but you will also be reported to the police. No banning involved, however it'll be quite hard to share your files when your computer has been impounded and you are in prison. ;-)

  • What problem are you solving? (Score:4, Interesting)

    by anthony_dipierro ( 543308 ) on Thursday May 30, 2002 @11:59AM (#3609776) Journal
    It seems like you're merely making minor improvements to something that's already out there. That rarely works, and I highly doubt it will work in this case. Your network is not anonymous, and you threaten to report IP addresses to the police. So basically your network is only for legal files. That's a fine niche, but what advantage do you offer over http and ftp? Probably not very much.

    • This work doesn't solve any new problems and is essentially the same as Chord [mit.edu], a project at MIT, which has the same basic layout but in a more structured fashion (as far as I can tell from his slides). Chord came out over a year ago (they submitted to last year's SIGCOMM which would have been due in early 2001). He mentions that they're very similar, but as far as I can tell, there is nothing new in his implementation and it isn't necessarily as good. The Chord guys actually prove how fast their system works while he just waves his hands. They also have a paper about how to actually implement a p2p file system over it (I think they give a reference to someone who actually did it too).

      There's no good reason this work should have been accepted. Whoever reviewed for this linux.conf.au dropped the ball in a big way. A real academic reviewer would have eaten him alive it.

  • Compatibility. (Score:3, Interesting)

    by Matt2000 ( 29624 ) on Thursday May 30, 2002 @12:21PM (#3609947) Homepage

    You're probably going to have to offer some form of compatibility layer with the other networks. That's how Limewire got it's improved Gnutella protocol out there.

    That is assuming that your protocol really does offer tangible benefits and people will want to continue using it when the other networks are available to them as well.

  • target a genre specific music community (Score:3, Interesting)

    by spaceling ( 521125 ) on Thursday May 30, 2002 @12:24PM (#3609968) Homepage
    You could popularize your app by going to a specific music community like ambient or idm at hyperreal.org and offering your application for file sharing among the list subscribers. Soulseek is a file sharing app which became popular for idm listeners (see soulseek.org [soulseek.org]). The great concentration of interesting files for like minded users will make your app competitive for these users. This user base should help you develop your app. Eventually the apps popularity should add more and more users from outside these smaller communities.
  • While you are trying to do something rather valiant, the problems arise in your contradicting statements:

    "...no censorship, no entry taxes, no one booting you off the network, and no weak point which can break the whole system."

    versus

    "In particular, if you try to share snuff or child pornography, I will be able to work out you IP and from that your location and identity. And I will report you to the police."

    So which is it? Censored or not? While you may be trying to do the moral thing here, this 'censorship' will be the downfall of your P2P network and it will probably not progress beyond your own little 'circle'.
  • could you fix one of the bugs in the old one? I tried to post an "Ask Slashdot" the other day about getting 3000+ emails from a misconfigured router that happened to get an ex-Kazaa DHCP address and thus was bombarded for several hours at a pretty impressive rate on port 1214. Any chance you could try to make the Circle route around dropped nodes a tad faster? I shudder to think what would happen to a poor 28.8k user if they had that happen to them....

    Also, after that experience and sniffing port 1214 with netcat for a bit to see wtf was going on, I'd say you might want to do some traffic analysis of other P2P networks if you're gonna try to censor the kiddie porn - you might be biting off more than you can chew, assuming anyone did use it.

  • So my question is this: how do things like these overcome network effects and become popular?

    One word: Content. Simply ensure that your [yet another] P2P network has tons of 1) porn, 2) pirated a) software, b) music, c) movies.

    You owe me $42,369 US in consulting fees. Payable immediately. Ping me offline regarding the bank account for the wire transfer.

  • Before the masses are going to make a switch the interface has to be extremely easy to install and use.

    Telling the average user he needs to install putty and and put plink in his path is like speaking japanese.The buety of limewire and bearshare is the user installs the software with one click and it works on just about any network.

    Also, making the interface attractive and bug free probably would'nt hurt.For example your windows client seems to have a bug where it keeps asking for the hostname of my firewall(I dont have one) and wont let me exit the program.

    If you write better software they will come.

  • Mainly you should use the network to propogate yourself. e.g. provide patches for the next version of Apache so that all of the users of the Apache web servers end up talking to each other and download from each other, rather than the server. This could beat the slashdot effect to some reasonable extent- a lot of files are fixed; particularly distros, jpgs, mpgs, html etc.
  • get yourself the new improved Slashdot Effect
  • other networks. Povide a conduit between it and gnutella and WinMX
  • I would argue that calling it the "Network Effect" is wildly misleading. Just because an orange grows on a tree does not make it an apple. And this effect extends well beyond network-centric developments.

    Why do people have different priorities for their choice of operating system? Why do people have different priorities for choice of monitor?

    If you want the real answer to "why", ask yourself, and I ask non-faecetiously: Why did you not adopt and support an existing P2P protocol instead of writing your own?

    Sometimes people adopt something because they share a belief system with the author - qMail is a good example. Other times it is because they have a need the authors are catering for - exim is a good example. Whilst other times it is because "better the devil you know" - sendmail is a large example.

    You could ask why the Roxen webserver (which predated Apache in its incarnation as 'Spinner') is virtually unknown - when its XML-like server-side scripting system, its performance and sophisticated caching systems would seem to make it todays ideal candidate. (http://www.roxen.com)

    In today's environment, however, the answer is usually that there are too many choices and not enough differences. We don't have time to do a complete study of all the options for each and every program for our requirement. We look at the first paint program that supports 'dongleberry' format and learn our way around that. We try three 'ldap' capable mail systems and go with the one that installed man pages not tex files that Wouldn'tBuildHere(TM). We choose our programmers editor based on the "summary" in the first results page returned by google, and we choose our P2P network by asking Harry where he got that awesome new Britney Strips video.
    • Network effects are a pretty interesting market failure (well at least to economists, I suppose others could seem them as about as exciting as watching paint dry) they are specifically any item that gets more valuable to users as more people use it. It is one of the only true barriers to entry which are costs that the initial user doesn't have to pay that later competitors do. This is the reason all the dpt coms wretched cash trying to gain first mover advantages from network effects. It turned out that for most of them there weren't any network effects and the money was largely wasted. We'll probably be living down those mistakes for a decade in some sectors (telecommunications).
      More specifically the interesting cases are items that are natrual monopolies, operating systems and telecommunications networks are the best examples. It is prohibitavly expensive to build another telephone network since you will not make enough if you only get 50% of the customers (I made that up on the spot), however two more competitive firms would be much less profitable than one single firm and consumers would be worse off imagine if you couldn't call half the phones out there.
      Operating systems are another less clear example, the more developers and users an OS has the more valuable it is since there are more useful applications, which are made because of the larger userbase. Other pieces of software are influenced by network effects, applications that allow additional application functionality browsers, appication servers, some servers, etc. and p2p because the draw is in how much stuff is there, think about how much simpler things were when Napster was the network over ftp/irc/usnet.
      Other examples are standards, which become more valuable as more people sign on to the standard. This is why Rambus' actions were reviled by JDEC.
      For most industries you are right people choose the first solution that works and gets their attention, however in industries with high levels of network effects the first company there usually has the advantage.
      • And that may hold true for schools now. But it doesn't cover all fields or realms. Where it tends to work is in tools that people use, rather than "toys" (no offense) like P2P networks.

        Many strange products that should have died early in their lifetimes have lived long and prosperous lifespans because they managed to get their foot in the educators door...
  • There is no competing P2P software on the Macintosh. MacPython [www.cwi.nl] is mature. Announce your port on www.macintouch.com. Instant 5% of all computer users.

Slashdot Top Deals

Understanding is always the understanding of a smaller problem in relation to a bigger problem. -- P.D. Ouspensky

Close