comments

Just wanted to note that I just recently registered pownce.biz (for powncing on business ;) -- in case anyone might be interested in developing a microblogging site, I think that might be one avenue that I would consider (and I'm also open to other suggestions)....

Essentially, this is a "numbers" game. If you've got the entire Internet on your homepage several times a day, then you have such large numbers that you could create a far more sophisticated system than if you have "only" 1% of the Internet or "only".01% of the Internet (I think digg.com boasts something 30MM uniques per month?)

So if you've got a homepage with extremely wide reach and a box to enter 1 or 2 words, then you ought to be doing something fantastic. If instead, you only do something mediocre (of even something inferior) then you will ultimately fail (for example, I believe it is not at all far-fetched that Google will soon implode due to engaging in "self-promotion" [blogspot,youtube,knol,etc] rather than providing at least "passable" results [as it is today, Google only provides passable results for "navigational" searches -- i.e, people who do not yet understand that they do not need to type "shoes" into Google to get to shoes.com]).

Oops -- went off on a tangent again... :O

If you only have a relatively small group visiting a page (to rate it) then you are perhaps not reaching "the law of large numbers" scale -- and so this is why you should choose your statistical method with a view to what can nonetheless provide meaningful data.

A Yes/No question is one thing, but asking a question may channel the responses into one particular pigeon hole. For example: is a humorous (or satirical) rant useful or not?

I can't remember what I wrote last, but I would (continue) to argue for a three-state logic: "approve", "disapprove" and "other" ("all other cases") plus a write-in box limited to one word (no spaces or hyphens) where people could write in "useful", "funny", "understanding", "knowledgeable", "informative" (and so on [note that this, although it seems like a quite "open" or "free" response is nonetheless focused on "predicating" the contribution that has been made]).

:) nmw

It's not a rule -- just an observation: I recall a statistic that more than half of the terms typed into to search engines are 6 chars or less (total query length!). I have no idea how any "rule" got started -- but then maybe I started it! ;D

I think if there is any myth, then that is that ".COM is king" -- investing irrationally exuberant sums of money into a domain name that is actually just a long string of letters (and not a meaningful term -- i.e., one that is sought after) is going to be very difficult to "pay off". It is far wiser to invest in keywords that people actually search for -- and there the choice of top level domain ("TLD") is a far more complex matter than simply saying .COM is the only TLD that matters (this ludicrous opinion is based on no data whatsoever -- and it is quite ironic that .COM is actually one of the riskiest TLDs on the globe -- northern Europe cc TLDs are far more secure [see e.g. http://gaggle.info/post/65/surfing-northern-europe-is-safer-than-following-american-links-online]).

The most vivid, memorable and relevant strings on the globe are English words: e.g.: cars, homes, hotels, pizza, etc.

I'll make this one short. :)

Today I updated my Facebook status to read:

>> Norbert thinks digg.com should allow its users to recategorize news that was submitted in the "wrong" category (according to popular vote). <<

Do others see how adding descriptors/tags to the thumbs up/down would be similar to my proposal for digg.com? Why not create a better system than digg.com? For more inspiration, here's a quote from Marrianne.COM:

Our deepest fear is not that we are inadequate. Our deepest fear is that we are powerful beyond measure. It is our light, not our darkness that most frightens us. We ask ourselves, Who am I to be brilliant, gorgeous, talented, fabulous? Actually, who are you not to be? You are a child of God. Your playing small does not serve the world. There is nothing enlightened about shrinking so that other people won't feel insecure around you. We are all meant to shine, as children do. We were born to make manifest the glory of God that is within us. It's not just in some of us; it's in everyone. And as we let our own light shine, we unconsciously give other people permission to do the same. As we are liberated from our own fear, our presence automatically liberates others.

I notice that after my "massive" brain dump of several days ago, that I overlooked to follow up on some of Karrie's questions. First and foremost: No I have not seen anything like the level of sophistication I describe here. I was quite involved in a similar community rating scheme developed with input from the community, but the archives from that project have been taken down from the Internet. To cut to the chase, I think this would be a "pioneering" scenario.

I have also run the the ideas through my "database" perspective -- and I must say that it is indeed very complex -- especially if some members are permitted to have "personal" tags (so I would perhaps limit this quite strongly).

I would say that the crux of the "tagging" system (I am going to stick with this term -- the descriptors actually function quite similarly to "tags"... indeed, so similarly, that it might be useful to think of the data architecture in similar terms [but that need not worry us here]): let's say the "community wide" vocabulary contained, say, 20 - 40 terms. For "information retrieval" (aka "search"), this would be the number of "types" the search engine would have to check, tally up into a score, etc. -- resulting in a rating such as "very funny" or "quite insightful" or "super helpful" (in these examples, the terms would be "funny", "insightful", "helpful"). IMHO, there is no reason why the same principle might not equally apply to such descriptors as "marketing", "accounting", "teamwork" etc. (but it might nonetheless be useful to differentiate between articles that are funny / helpful and articles that are about how to be funny / helpful.

Here are some more concrete suggestions -- I would suggest three "levels":

0: Basic membership

allows members to choose e.g. 4-6 terms to tag from palette of community-wide terms

1: Active membership

allows members to choose e.g. 6-8 terms to tag from a palette community-wide terms (terms may be "freely" assigned until the community-wide vocabulary is "filled")

2: Supporting membership

allows members to choose e.g. 6-8 terms to tag from a palette community-wide terms plus 1 freely chosen term

The availability of freely chosen terms means that the language will predominantly be shaped by active+supporting members, but every member will still be "free" to apply his/her chosen terms as he/she feels is most appropriate.

Some might find it "quirky" that it might be "necessary" to have 2 levels of vocabulary (community-wide vs. selected terms). What is really going on here is that this helps members to focus on what they feel is most important from their specific expertise (or "point of view"). So whereas I personally may not be an expert at deciding whether something is "world-changing" or not (and perhaps I might not care, either), but I may instead look for whether something has the attribute "promotes social cohesion" -- and if that is something I feel strongly about, then I might choose that as an attribute to "rate" things. So the point of 2 levels is to allow what might be referred to as "linguistic relativism" -- like Eskimos have many words for "snow" vs. Germans have many words for "beer".

Well, I guess that all sounds quite abstract (and I guess it is abstract).

One simple approach to "get the show on the road" in a simple / straightforward manner would be to simply start out allowing every user to use up to 4 tags and the community-wide vocabulary would "fill up" at say 40 terms. Then, at that point, only the 40 most popular terms would be maintained (and if someone had selected a term outside of this vocabulary, it would then be set to "undefined", "nondescript" or something like that).

This simpler approach would perhaps be a way to "kickstart" such an approach without introducing alot of complexity. Later, this could be refined to add such "bells and whistles" as providing methods for allowing the language to "evolve".

OMG -- I didn't want to write this much! :O

Sorry again -- I will have to practice being more succinct (if at all possible?)...

:) nmw

Yes, agreed that mandatory comments are neither "user friendly" nor "easy" to execute.

But having the system keep track of who is voting for what is both user friendly and relatively simple -- and would also be very informative to boot.

Joe said:

"So, the upside of your argument = more responsible, more positive feedback."

In my first response I noted that it is very important that feedback be transparent -- such that anyone (or at least any member) can tell who voted which way. If all votes are carried out by phantoms, they aren't as meaningful as if someone is willing to stand behind their own opinions. There is a certain fanaticism supporting "secret ballot" voting -- but I strongly doubt that secret ballot voting would lead to reliable (or to use Joe's term: "responsible") results.

Yes, I agree with you, too, Judy! ;D

Indeed, if you recall the suggestion I made about a week ago it in fact had much more to do with personal preferences -- namely enabling users to find things via the help of people whose opinions they personally value.

Indeed: There are many ways to find "relevant" information -- and it is wrong to think that one way is the best for everyone every time.

Bob,

seeing as I have just written so much, I will take it that I am being addressed when you mention "obsession" -- and it is indeed probably true that I am obsessed with "separating the wheat from the chaff" (insofar as that is my profession). And I do not expect everybody to have this obsession -- others might be "obsessed" with art or writing or marketing or sales (I guess everyone focuses on what they're interested in).

The question for people with a limited amount of time is: If there are 100 articles about a topic (e.g. "successful marketing") is there any way to help a person interested in that topic to decide which one to read first?

Of course if there are only 1 or 2 articles on a rather particular topic, then it may indeed be much more effective to "search out" the detailed topic (and for that I would agree that it may be good to look to other methods of finding the most appropriate information [such as -- for the above example: "which articles do successful marketers recommend?"])

Karrie said:

"However, I do not like that if a voter’s descriptor is changed that it would delete that vote for all those articles in which that descriptor was used to vote—this could hurt the article author who wrote a great article."

Yes, I agree -- but actually the vote wouldn't be "lost" -- just the descriptor.

I also think that the numbers I suggested above may make it sound more complex than it actually is. I expect that I personally wouldn't need many descriptors (over time, I have found that "insightful", "interesting" and "funny" are most useful -- perhaps "well-written" might also be a good one to start off with).

Let me give a concrete example. All the paying members start off by voting for something and start assigning such descriptors as "funny" or "well-written". If say 1% of the membership uses the term "funny" (or maybe simply start off with 10 people or so) then that term becomes available as "community-wide" term to rate content with.

Note that in order for such a system to be useful, these terms need to stay more/less fixed over time. If a person could say something is "helpful" today, but next year "helpful" is dropped and "useful" is used -- does that turn what used to be helpful into something which is now useful? No, not necessarily -- and this is why when a descriptor is dropped the "qualitative value" should be transferred to nondescript.

This system is still far from perfect -- I may find something very funny but not at all helpful. But implementing such a system would involve something like 3-state logics (positive, negative and undefined) -- such that I could assign "funny" +1, "helpful" -1, and everything else 0 -- now that is complex.

To keep it simple, I agree that only "positive values" should be considered -- and that would also avoid any argumentation about "negative" sentiments.

So let's say that each member could press a button that would open a box saying something like "I appreciate this because" and then 8 checkboxes appear. The user could focus on one aspect and therefore send a clear and distinct message, or the user could distribute their appreciation over several (or even all aspects).

Then how would such ratings be presented on the site?

You could say "most appreciated" (in general -- simple tally of all the people supporting), or "funniest" or "most useful". I think, however, that the precise method chosen to present the information is rather secondary -- the primary issue is to first decide which information to collect.

:)

LOL -- no, I'm an information scientist (and designing such systems is what information scientists have thought about for many many decades) -- I guess Melvil Dewey might be considered to be one of the "first"! ;D

( BTW: I wanted to add a ps [but was too late]: perhaps with a better rating system, it will no longer be necessary to have "requirements" [e.g. length for articles] -- whereas a poor article that might meet certain criteria might then squander in some old dust-bin, a very good article that happens to not fulfill certain criteria may nonetheless be promoted by being highly rated. :)

Both Karrie and Pamela present very good ideas (and I have been an adamant proponent of such systems, too). Before I elaborate, let me add that it is really important to be able to link (point) to what is being rated -- and in some cases this may even be not a text, but rather a person (for example: when the person says "I'm sorry" and/or acknowledges a mistake). In addition, comments to an article should also be "rateable". And the article itself and the discussion ensuing from that article should be rated separately -- and it's primarily the discussion that should allow revisions of ratings (for the obvious reason that a discussion is open-ended).

Now to the particular case of "descriptors" or "tags" which could be used to make a qualitative (rather than a quantitative) statement. I would suggest that each member of the community be allowed to have a certain number (I like 8, because I like the fact that I have 8 fingers ;). And these 8 labels could then be applied as checkboxes to anything being rated. Anything that is rated would have descriptors listed in the order in which such discriptors have been assigned (in proportion to the total number of decriptors). If a person assigns 2 descriptors, each descriptor gets 1/2 wieght; for 3 descriptors, 1/3 weight (an so on).

Non-Paying members are not allowed to assign new descriptors. Then for paying members, there might be a limited ability to add new descriptors (say 3 or 4), and then at a higher level of membership all 8 descriptors could be created/modified at the whim of the member. When a certain number / proportion of members has applied a descriptor, then that descriptor can become a "community" descriptor, and then it is no longer the "personal" descriptor of a paying member. The number of community-wide descriptors should be limited to a couple dozen or at most a couple hundred terms (since otherwise the descriptor selection process becomes too difficult). If this certain limit is reached, the communty should vote which descriptors are superfluous and should be lopped off (perhaps a more sophisticated system would allow paying members to choose from a larger vocabulary of descriptors/terms than non paying members, but this may indeed be "overkill").

Note that in this system each person would be limited in the number of descriptors they are allowed to apply in general to anything (maybe paying members could be allowed to use 1 or 2 more descriptors in their "personal vocabulary").

In general, however: less is more. If I have to choose from 100 descriptors, I would probably not choose at all -- but with just a handful of options then it's easy.

If someone were to remove a descriptor type from their palette of descriptors, then all of the things that descriptor has been applied to in the past would lose this information (and in that case the "weight" would also be lost -- for this purpose the user would be forced to have 1 of their palette of descriptors set as "nondescript" and the weight of the previous descriptions would be transferred to this "nondescript" tag). Hopefully that would be sufficient to discourage people from assigning descriptors haphazardly.

It is important to note that this is something that the community needs to "control" (it is a referred to as a "controlled vocabulary" for good reasons! ;). Otherwise, such descriptors would become similar to the "uncontrolled" alphabet soup you might find at a place like del.icio.us or something like that.

Sorry if this has been a little technical -- but it is actually very important point in having a well-functioning community.

Kudos to the managers of the site for engaging in such an open discussion about such critical issues!

:) nmw

Oh, DRAT! :(

Here's the "corrected" link:

mp3 cast

I also added the "correct" link to the digg.com page I linked to above -- so now you can also find the correct link there.

Thanks for alerting me to my mistake, Howard!

:) nmw

Well, I too "understand" how to name things (;D), but I cannot give any kind of legal advice regarding trademarks. Also, I generally focus on search terms (dictionary words) rather than "names" per se.

Here's an interesting podcast of a SWSX panel discussion that I just posted at digg.com (might be good to listen to for some ideas):

2008 SXSW Interactive Podcast Logos Why They re Irrelevant

Here's a podcast of a SWSX panel discussion that I just posted at digg.com (might be good to listen to for some ideas):

2008 SXSW Interactive Podcast Logos Why They re Irrelevant

My experience with rating systems has been that they work better if people can see who has voted (and which way) -- and yet I know there are also vociferous adherents of "secret ballot" systems.

In a more "sophisticated" system, each user might be able to assign their own "weights" to arrive at "scoring" -- such that I might weigh "legal expertise" differently than "artistic expertise". Somesthing like this would probably only be feasible to keep track of in a small group (such as "my network").

So there could be 3 levels of "scoring":

1. biznik-wide

2. my network

3. according to my weightings (and perhaps there could be 2 or 3 alternative weightings)

Perhaps levels #2 and #3 need not be calculated for every page-load, but rather with a mouse-over or a click (in order to reduce the computational load).

:) nmw

Oh, I'm all for niche.names -- indeed, I'm into all sorts of .names!

:D nmw

BTW: I find urbandictionary.com to be a good resource for such slang (and/or "new" terms):

solopreneur

LOL!! :D

It's not so important how often a word is written -- what's more important from the point of view of being found is which words are being searched (step 1 of the 2 steps).

Then, when you are on the list of "results" for that keyword, it's important to motivate people to click.

Of course that's all IMHO.

BTW: I happen to know a NO survivor who's very big on the "resiliency planning" concept (note, however, that if you keep talking about it long enough, there may actually be someone who might just snatch up a domain from right under your nose! ;)

Hi Viv,

I think if you are trying to coin a new word, then you are taking on a large battle. Granted, you're not saying you want to coin a new word, but rather asking whether X is an "up and coming word".

For the sake of argument, let's pretend that "solopreneurs" is such an "up and coming" new word.

1. Do you expect someone to go to their computer and say "I want to find out all there is to know about solopreneurs" or maybe "where can I find a solopreneur?"

2. If yes, then perhaps the name might be good for your expectations; but I actually doubt it -- I expect that less than a handful will type in that term in a year (note that I also acknowledge that very few would type in "disaster planning" -- simply because they're human and would rather plan for "success")

3. I think you need to look at this issue as a process with at least 2 steps:

3.1. search = "where can I find a list of X?"

3.2. selection = "now that I have a list of X, which one should I contact first?"

Domain names for 3.1 (e.g. cars.com, hotels.com, pizza.com, etc.) are exorbitantly expensive. In contrast, domain names for 3.2 might be freely available to register.

I wrote an article about these "general" issues yesterday, and I don't want to get off the topic of which name would be right in your case here. The general article is at "the difference between a brand name and a domain name".

see also this article on businessweek.com and/or my comment in reply to the story.

BTW: Google is no longer very good at identifying valuable content. One reason why Google was good in the early days of the net is that it was mostly university professors/researchers who were writing back then (see also the Wisdom of the Language)