ietf-asrg
[Top] [All Lists]

Parameters for success? (was Re: [Asrg] Re: [OffTopic - NNTP]

2003-03-23 15:52:52

On Sunday, March 23, 2003, at 10:16  AM, Kee Hinckley wrote:

I have never understood the resistance to using newsgroups. Yet customers choose web boards (which I despise) over them every time. I think it's just a question of what people are used to. The came to the internet via the web and then discovered email (or vice versa). Now you want to introduce a third piece of software, and they've already got something that manages to do solve the problem so-so.

to some degree it's a generational gap (not necessarily by calendar age, but by how long you've been on the internet). you see these all over the place on the ent -- HTML email is another one folks fight over, coerced reply-tos on mail lists is a third, and increasingly, IM vs. email. the new generation of users is more of an IM first, e-mail as backup, where us old pharts tend to live in e-mail and remember fondly the days when NNTP was the primary dinosaur.

Which pointed out to me something I think this forum ought to be keeping in mind, the issue of parameters for (and barriers to) success. Whatever the solution(s) that we come up with, there are certain things that can keep any solution from succeeding, so I think perhaps it might be reasonable to think about what those parameters and barriers are.

Some of the most obvious things that come to mind for me are:

1) e-mail is easy to use. The more we complicate e-mail to solve this problem, the less chance joe typical user will agree to do it. If you don't make the benefits cleaerly visible to the user and the ease of use reasonable such that my mother and our aunt sally can (and will) accept the solution, it's a non-starter, the way PKI and S/MIME effectively are. Those things exist, but basically, nobody but geeks use it, and even most of them don't, because the hassles outweigh the benefits. (think of this as the 'seat belt' problem. almost everyone agrees we should wear our seat belts while driving. A good percentage of people don't anyway, even though the hassle of wearing one is generally trivial. That's a social issue that has to be dealt with of the solution won't succeed)

2) It has to save a person/company/ISP/organization/etc money, time, hassle. If it doesn't, it won't be implemented by that person/company/ISP/organization/etc. show me the money, basically. "for the good of the net" will get the EFF to install it. will sony? IBM?

3) If it requires implementation support from large organizations or ISPs, any solution that doesn't have buy-in and a hard date forimplementation from AOL is a non-starter. To a lesser degree also true of Yahoo, MSN/Hotmail, Earthlink, wannadoo, and maybe half a dozen of the larger international ISPs. You don't need ALL of those, but without AOL, you might as well go home, and until you get half a dozen of the top 12-15 ISPs or organizations, network providers, etc, you don't have a chance at building critical mass.

4) will it scale through organizations? I've seen a number of suggestions that really come across as if the proposer sees all e-mail environments as variations of THEIR environment. A solution that isn't tied to the individual (which arguably it can't be, since that solves none of the network bandwidth problems) has to scale from my site (a box on the end of a DSL line serving 6 users, 15 mail lists and a couple of hundred messages through those lists a day) to ibm.com or microsoft.com, to MSN, to yahoogroups, to hotmail, to AOL. If the solution can't scale up OR down to fit those needs, it's going to be a hard sell for success.

5) will the senders buy in? If not, isn't it a non-starter again? Not just mail list owners like me, or list sites like yahoogroups, but the large-volume mail senders, whether it's content (cnn.com) or e-marketing? How do you make it work without them? How do you convince them to join in?

Other thoughts?

I've found myself, as I've been watching stuff sent back and forth and discussed, asking myself things like "how would this affect my systems?" "Does it make my life easier or more of a pain?" "does it save or cost me money?" and the biggie: "How would we ever convince AOL to buy into this?"

think about AOL's position right now. Take, oh, e-stamps, or sender pays. how does AOL implement that on its environment? convince it's users to use it? migrate its users off of existing email? How long would that migration take? What if they weren't willing to change or don't upgrade their software? 25 million users, 3-5 billion pieces o' email a day, and users who haven't upgraded their software since they plugged in that windows machine 4 years ago, and won't upgrade any time soon. How do these plans work in THAT environment?

And if you don't get AOL bought into it, given their size adn volume, can you make it work without them?

I realize it's way early to get into too much detail on implementation issues -- but understanding certain necessary barriers to success will allow us to NOT go building things until we know those barriers are taken care of. If, for instance, it requires server upgrades by AOL or UI changes to the AOL client, for instance, I think it's silly to go beyond a certain level of discussion until someone finds out if AOL (or hotmail, or yahoogroups, or....) are willing to consider it. If they say no, it's better we consider lesser solutions that don't require their cooperation than building a better one that'll fail because we can't convince them to do it...

To me, that implies a hierarchy of solutions that can help us set priorities and understand the complexity and level of cooperation needed for success:

on the receiving end:

1) individual: build something an individual can download and implement. advantage: easy to do. disadvantage: limited ability to manage bandwidth issues, maintenance and user capability. (example: Apple's Mail.app junk mail filters)

2) site: things an admin can install to benefit the users of their machine. advantage: fewer people need to take the action, can affect network bandwidth issues as well as individual mail boxes. disadvantages: loss of user configurability (perhaps), loss of customization. (example: spamAssassin)

3) organization: things an organization can do to implement solutions. further down the scale of size from site, but effectively the same issue, scaled to multiple mail servers and a larger user population, plus perhaps issues of multiple locations and nationalities, meaning different parts of the system, might be regulated by different legislation.

4) ISP: seller of services and bandwidth to organizations. (for the purposes of what I'm saying, hotmail is a large organization, UUNET is an ISP, SpeakEasy is depending on which services you use, some of both)

5) infrastructure: making changes to the underlying structure of the internet (root name servers, protocols, etc) and the software that runs it. Basic, fundamental changes to how the net operates. (rDNS, new protocols for authentication, etc go here).

A) stuff that doesn't cleanly fit in the above, like yahoogroups, large volume senders, e-marketers, and stuff I'm forgetting off the top of my head...

The further you move from the individual level, the harder it'll be to convince everyone to implement, but the more chance you have of significant impact on the problems. So I think it's useful to at least casually keep in mind where specific proposals fall, so we can scope things like complexity of implementation and what buy-ins we need to get before we have a realistic chance of success.

(I think it also is a good visual reminder of why we need suites of solutions; indvidiual and site solutions to bridge us towards the longer-term, more strategic and structural ones -- and who needs to buy off on those solutions before we can hope for them to work)



--
Chuq Von Rospach, Architech
chuqui(_at_)plaidworks(_dot_)com -- http://www.plaidworks.com/chuqui/blog/


_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg



<Prev in Thread] Current Thread [Next in Thread>