time for article review process?

Posted 1 Jan 2004 at 23:34 UTC by lkcl Share This

a recent article on advogato leads some to believe that a "review" process should be implemented. here are some of the reasons why that process may not work, and some alternative recommendations.

The issue is that some people believe that material posted on this site may not be of interest to all people who read here and in the future: that is a subjective opinion of theirs, and the charter of the advogato site is much broader than their opinions. i have to say that whilst that may be the case, personally i genuinely look forward to improvements in the information quality of advogato.

the crux of the matter is that advogato's charter is broad, even within its specialisation, and the expectations of the specialised advogato readers is polarised, as might be expected of any diverse group of people.

opinions differ: that's life.

so how can the quality of advogato's information, even though it is a specialist site, be improved?

there are two ways:

1) filter the posting rights so that people who are not considered, even within the specialised community, to be sufficiently responsible enough to provide decent information, cannot immediately post.

i.e. even though between one and one hundred people certify them as being trustworthy, they are still not considered to be trustworthy _enough_ to put information before others without some sort of review process.

2) filter the content rights so that people themselves may choose which information and which topics they personally are most interested in.

i.e. like the diary rating system people choose a threshold for their interest level, and others provide opinions in certain categories.

now, how might 1) - a posting rights review process - actually work?

well, presumably, it is considered that a certain number of "masters" should review an article and "Certify" that it passes muster: such a system was implemented in http://sf.net/projects/virgule, where Articles themselves could be Certified. "Trust" flowed from the top-level seeds (People), via People, and ultimately to Articles.

a simpler (but one that i feel is less elegant but it achieves almost the same thing in the end) system is to simply count up the number of Masters who set an "approved" stamp on the article. [the difference is subtle: it's possible for the "Metric Flow" to peter out by the time it gets to Articles if certain People have not been Certified enough, themselves. it's to do with flow and degrees (the number of links between you and the top-level seeds) and in the former system, Articles are one more degree away from the top-level seeds].

the trouble with both these approaches is this: that people may create fake accounts, Certified by themselves, with which to "approve" their own articles. The creation of "fake" accounts could go several layers deep.

At first glance, a solution to this particular problem appears to be to back-track the Certification process to see if there are any bottlenecks of only one individual: the trouble is, how do you distinguish between a real group of people that are Certified, ultimately, by only one person, and a fake group?

Answer: you can't.

There is the distinct possibility that you will end up cutting off legitimate article writers, and they won't understand what is going on, and only someone with sufficient technical knowledge will be able to explain to them why. unless of course the Certification tree is automatically analysed and an automated explanation provided as to what is further required for the Article to appear on the site.

The answer to that problem is this: that in order for a Person to be Certified, they require three separate Certification paths (not just one).

In other words, if you believe that the solution to improving Advogato's informational quality is to have a multi-person "Review" or "Certification" process on Articles, then you must also have a multi-person Person Certification process to back it up, otherwise a multi-person Article Certification process is useless.

Raph and I discussed, two years ago, implementations of multi-person Certification paths. Raph believed that the implementation was simpler than I understood it: I was unable to communicate enough and understand enough to be able to explain why I thought he was on the wrong track. My inability to explain is in no way a failure on your part, Raph: I feel that it is important that I emphasise that to you.

Since then, I have thought of an alternative implementation solution, one which involves pre-analysis of the tree before the Net Flow is applied. There are two possible solutions, and they are both based on width-wise tree walking instead of the depth-first Ford Fulkersson Maximum Flow algorithm.

The present net_flow.c implementation performs a depth-first pre-analysis, and this stage could be adapted simply to remove any nodes at each level of depth, to not walk down any nodes that do not have more than three incoming Certifications. Given that the depth-first analysis process BEGINS from the top-level seeds and proceeds ONLY to those people who are Certified from those top-level seeds, the added restriction of only walking to the next depth level down nodes that have three or more incoming Certs has to work - as long as the analysis is forward-looking, and Certifications that are not linked back to the top-level seeds are pre-pruned.

To clarify and summarise: if the graph is restricted to include only those people who are linked to the top-level seeds, and that there are always three or more independent paths leading from each person on the graph back to the top-level seeds, then we have a workable system for doing multi-person Certification.

[Hey, it makes sense to me: I realise that the above paragraph may not be entirely too clear as i've tried to over-explain - sorry :) ]

The second possible implementation is to combine the above approach with an alternative Maximum Flow algorithm, one that is depth-first in the first place.

So, to conclude the analysis of the first approach - namely to put in place a review process - it is necessary to extend that review process to People. In light of the weaknesses of the present Certification system, adding in "Multi-path" Certification of People is a definite advantage.

However, even if such a process is implemented, it still doesn't solve the problem of many like-minded and real people "straying" in some way from the "ideal" - or just simply _different_ - interpretation of the site's charter, and in the end, if you raise the bar to four, give, six people, you just end up putting people off from using the site.

which brings us on to the other alternative 2): user-driven interests and ratings.

The principle is simple: apply a slashdot-like category for the article, which can be made by the writer, and then other People can "rate" the article in that category with a value (in the same way that Diaries can be rated on Advogato at the moment).

People can then select a threshold of interest in each of the categories for articles, and then articles below that threshold disappear from their radar.

A slightly more advanced version of this approach is for People to be able to rate the Article independently in all categories, not just the one that the author believes the Article to be in: the author's opinion of what slashdot-like category the article should be in is reduced to an advisory capacity.

The advantage of the advanced version is that it takes into account many opinions rather than just that of the author.

In order for this approach to be useful, it is important that it be tied in to the Metric system. One such way to make a tie is to only use the "Ratings" of People with Certifications above a certain level (e.g. Journeyer) - or for the "Rating" to be "multiplied" by the Certification level as a number. This approach gives Masters Ratings more weight.

The whole point of using this approach is to give readers better information quality, and they can do that by selecting top-level categories. Not only that but they know that when they review articles in those categories, they stand a good chance of actually reading something that will actually be worth reading.

An advogato inside an advogato.

This is, in my opinion, the only way in which Advogato is likely to be accepted by such a diverse group of specialists: to categorise articles by further specialisation, just like slashdot does.

The important thing is to do that categorisation in the same spirit that Advogato follows, not in the spirit of slashdot.

So, on the one hand we have one proposal to restrict information, and on the other hand we have another proposal to further categorise but not restrict information.

The spirit of advogato has always tended towards making positive statements by NOT making negative statements.

The conclusion is, therefore, that categorisation and further specialisation would be a better way to help Advogato readers' see what they want than it would be to attempt to stop writers just because a small sub-group of people don't believe that a certain article fits the charter of the site.

The proposal to add categorisation allows people with clout to take a positive lead in getting better information to Advogato's diverse readership.


Too much and too little, posted 2 Jan 2004 at 13:47 UTC by chalst » (Master)

It seems to me the problem is as much people not posting enough good articles to the front page as too much crap appearing, so adding categories of content risks fragmenting the little good content we have.

On the other hand, categories may encourage people to post content they would otherwise fear has too narrow interest for what now is an article. I've thought that creating something like the ACM Special Interests Groups for advogato would be a good thing, which could have their own root pages.

On certifying articles: I often think the right thing is just to get someone to criticise an article: there's a lot of flawed ideas posted that could have been good articles with a bit more thought. So I'm more for a refereeing process than a certification process; heaven only knows how that would be organised.

It seems simple to me, posted 2 Jan 2004 at 16:18 UTC by jds » (Journeyer)

The front page of Advogato will only display articles which have been approved by a percentage of the number of Masters who have posted diary entries in the preceding thirty days. This way, the number of Masters is not hardcoded, but simply dependent on how many are using the diary portion of the site (and therefore visiting the site more often than once in a while). Also, which specific Masters are involved is not predetermined.

A secondary "pending front page" page can be posted to by anyone who presently has posting rights. The more diversity that appears on this secondary page, the better! Thus, the outside world will only see the unified front of Advogato. Those who dig deeper can find the whole while beautiful spectrum that has currently been Advogoato's front page.

In this way, those geeks (we've all been or known one) who currently use the attention-grabbing technique of "posting to the front page" as a learning tool can continue to learn, without garbaging up the front page. I personally have refined my thinking process based on responses to a half-rant article posted on Advogato, and am grateful for that.

chalst: "there's a lot of flawed ideas posted that could have been good articles with a bit more thought." Well said. Important to note that this can be taken too far (articles written by committee can be impotent).

lkcl: this present article will probably not make it to the proposed new front page, while containing sentences like this in it: "[Hey, it makes sense to me: I realise that the above paragraph may not be entirely too clear as i've tried to over-explain - sorry :) ]"

"The spirit of advogato has always tended towards making positive statements by NOT making negative statements." I'm glad someone put it into words. Thanks.

Much simpler solution, posted 2 Jan 2004 at 21:16 UTC by pphaneuf » (Journeyer)

The two losing articles currently on the front page were posted by "observer" level accounts (most probably the same person). I thought you needed to be certified to a minimum level to be able to post article? Apparently this is not (really) the case.

Just make this so, and it should already be better than it was.

Also, the previous losing article (the one that was a question about RPM, a little while ago, check the archive) had been posted by an "apprentice" level account. I seemed to remember requiring at least "journeyer", which would have made everything just peachy.

In fact, I think "observers" couldn't even post replies to articles, only post diary entries, isn't that the case anymore?

It was an attack against mod_virgule, posted 2 Jan 2004 at 22:04 UTC by trs80 » (Apprentice)

raph indicated that it was an attack, presumably against some vulnerability in mod_virgule itself. For a brief while the attackers were rated as Journeyers, so they could post the articles, then the ratings were removed. Anyway, raph has said he's working on adding article approval (although I can't find where he said this).

pphaneuf: I'm only an apprentice, but I'd like to post articles at some point (when I get around to actually writing them) - approval to weed out the obvious crap is all that's needed IMHO.

A counter-proposal, posted 3 Jan 2004 at 12:41 UTC by chalst » (Master)

I think there are a lot of good ideas in lkcl's proposal, but I think it suffers from two problems. The first, which I hinted at in my previous post, is that I think it tackles the wrong problem, namely degree of authorisation. I think lkcl's idea of lines of authorisation (let's call a certification with n independent lines of authorisation a rank n certification) maybe can partly solve the problem with overcertification elsewhere, say if we demand that journeyers must be rank 2 certified, and masters at least rank 3 certified; my guess is most bad certifications would just vanish. The second problem is complexity: I'm sure we don't need to rethink the whole idea of trust-metric, with all the experiminatation and fine-tuning it requires, to solve this problem.

So my proposal is this: we create a list of named editors. Anyone can submit an article draft (now not just Apprentices and above), but the draft only appears on the submitters homepage, and on an editor-only visible queue. Any editor can look at the items in the queue, and accept, reject, or call for amendments. Accepted articles appear on the front page.

Determining who is and is not an editor is likely to be contentious. I see no reason why this responsibility should be restricted only to masters; both fxn and mikehearn would make excellent editors IMO. At first I don't see why raph shouldn't just choose 10 or so editors right off; a more democratic process for hiring and firing could be put in place later.

It's not just the spam, posted 4 Jan 2004 at 17:24 UTC by etrepum » (Journeyer)

No offense to anyone, but the "articles" as they stand are on the whole pretty low quality. There are plenty of "articles" that are actually just announcements. CodeCon 2004 had two asking for papers, Savannah asked for help (and then stated that they didn't really need it so much anymore the next day), FSF asked for nominations, etc.

Conferences need papers, large projects need help, and awards need nominees, but are advogato's articles really the way to do it? Wouldn't a blog entry do?

Personally, I'd vote down 90% of the "articles" based on spelling, grammar, and formatting errors alone. If a review system is put into place, it definitely needs to allow comments to the author for revision. I'm not even sure that masters should be allowed to post directly to the front page without any review. I'm not convinced that masters are better authors than anyone else.

Advogato Governing Council, posted 4 Jan 2004 at 23:55 UTC by mslicker » (Journeyer)

I think chalst is on to to something. Perhaps a carefully selected Advogato Governing Council will lead us down the path to a free and democratic Advogato.

Re: Advogato Governing Council, posted 5 Jan 2004 at 03:03 UTC by tk » (Observer)

Yes, freedom and democracy are the two greatest things in the world, greater than even food and water, aren't they? When there's freedom, great thinkers like nospamplease and nospampleaser will be able to enlighten us, the ignorant Advogato masses, with their profound anti-Microsoft philosophy, and badvogato can open our eyes to the non-programming world of Go, and the greatest programming in the world who programs by not programming can instruct us on the cosmology underlying a certain postfix language.

Re: Advogato Governing Council, posted 5 Jan 2004 at 03:03 UTC by tk » (Observer)

Yes, freedom and democracy are the two greatest things in the world, greater than even food and water, aren't they? When there's freedom, great thinkers like nospamplease and nospampleaser will be able to enlighten us, the ignorant Advogato masses, with their profound anti-Microsoft philosophy, and badvogato can open our eyes to the non-programming world of Go, and the greatest programming in the world who programs by not programming can instruct us on the cosmology underlying a certain postfix language.

Re: Advogato Governing Council, posted 5 Jan 2004 at 03:03 UTC by tk » (Observer)

Yes, freedom and democracy are the two greatest things in the world, greater than even food and water, aren't they? When there's freedom, great thinkers like nospamplease and nospampleaser will be able to enlighten us, the ignorant Advogato masses, with their profound anti-Microsoft philosophy, and badvogato can open our eyes to the non-programming world of Go, and the greatest programming in the world who programs by not programming can instruct us on the cosmology underlying a certain postfix language.

Re: Advogato Governing Council, posted 5 Jan 2004 at 03:03 UTC by tk » (Observer)

Yes, freedom and democracy are the two greatest things in the world, greater than even food and water, aren't they? When there's freedom, great thinkers like nospamplease and nospampleaser will be able to enlighten us, the ignorant Advogato masses, with their profound anti-Microsoft philosophy, and badvogato can open our eyes to the non-programming world of Go, and the greatest programmer in the world, he who writes useful programs by not writing useful programs, can instruct us on the cosmology underlying a certain postfix language.

democracy - ho hum, posted 5 Jan 2004 at 11:16 UTC by lkcl » (Master)

democracy results in people being led by leaders that they _deserve_.

democracy also serves to weaken a country by making it impossible to put in "people-unpopular" laws - such as in france where the economy is likely to collapse in the next ten to twenty years because of the imbalances between retirement age, working hours, pensions and the actual number of people in work. any government that put in mandatory pension schemes in an attempt to solve the problem would be voted out of power within minutes.

the most effective way to run a body / group is to have appropriate laws / rules in place where the people running the group _have_ to follow those laws / rules and it is therefore pretty much irrelevant who does the running.

the problem with having "editors" is that you are back to square one and have undermined the principles on which advogato is based.

yes, i implemented the means to store articles "pre-front-page" in http://sf.net/projects/virgule's example sites. yes, they required a "status" of "Draft", "Preview" and "Final". yes, they required "approval" by any amount of People.

no, that doesn't make it a perfect solution, but that at least solved the "think about it a bit more" problem, which is different from solving the problem of quality.

I did, however, implement jds's idea "by default" because anything with a "Final" status went on the Front Page, when you went to http://xmlvl.net/articles you got a list of everything with "Final" or "Preview", and when you were the Author of an article you saw anything in that list which you had set "Draft" status as well.

chalst, i recommended the capability to have Articles appear under _multiple_ "Categories" which avoids the issue of fragmentation: any Person could "rate" an Article in a particular Category just like a diary entry.

the mechanism by which this could be achieved could be done with Certification, easily enough, as well as "rating".

WE NEED SOMETHING.

if nothing else, just something to make sure that those people with a sense of humour failure can at least be alerted "oh, this one's in the humour category, ah well i don't feel in the least bit like being funny today, i won't bother looking at it".

Regarding specialist skills: just because people are good at writing code and/or documentation it doesn't make them appropriate people to tell the world how good their code and/or documentation is.

Everyone has specialist skills: the most common mistake specialists make is to believe that their specialism makes them good at doing everything or anything else. Far from it.

which brings us right back to the democracy thing: people in general and in quantity have absolutely no specialist knowledge about running a country. there is a fundamentally mistaken belief that volume makes right.

i remember posting to microsoft on some corrections to a TechNet article on samba and NT interoperability: some people were setting "domain master = yes" when they already had an NT domain controller: samba was therefore taking the NT domain controller out because you cannot have two domain controllers for the same domain. when the microsoft techie posted on the samba mailing lists asking for corrections and advice, dozens of people responded - with the wrong information.

my correct advice was swamped, and that MSDN tech report still contains incorrect advice.

microsoft themselves are suffering from the same problem, with the original experts who wrote NT being ignored by the thousands now employed to work on it.

p.s. the u.s. isn't run by democracy: democracy is only recommended by the "free" worlds in order to keep other countries in a weakened condition.

i would do the code, posted 5 Jan 2004 at 11:50 UTC by lkcl » (Master)

p.s. i would code up Categories if i knew that raph would put it on the site.

Advogato Governing Council -- First Appointment, posted 5 Jan 2004 at 13:50 UTC by mslicker » (Journeyer)

I'm pleased to announce the first member in the newly formed Advogato Governing Council:

Minister of Truth

Re: Advogato Governing Council -- First Appointment, posted 5 Jan 2004 at 15:13 UTC by tk » (Observer)

*ignores mslicker's unilateral undemocratic appointment*

By the way, replying to articles is still dog slow... if there's no way to even speed up posting of article replies on this site, it doesn't seem meaningful to talk about more high-level changes.

Conjecture, posted 5 Jan 2004 at 16:49 UTC by mslicker » (Journeyer)

  1. The first sign of forum rot is a self-appointed Minister of Truth.

Re: Conjecture, posted 6 Jan 2004 at 03:26 UTC by tk » (Observer)

Excuse me? Who was the troll who started the talk about a "Advogato Governing Council"?

And what's your idea of a "forum rot"? When people don't talk about Forth as often as you would like? Or when people don't extol the virtues of freedom and democracy and Marxism every step of the way? Or when Advogato doesn't follow every one of your whims and fancies?

*sigh*

Need material to review/filter, posted 6 Jan 2004 at 14:42 UTC by DeepNorth » (Journeyer)

I hate to say this, but there was a time when I had two or three of the articles on the front page posted over a period of days. Fact is, there is not enough material being posted (or wasn't) to even worry about reducing or fragmenting the articles.

Popular sites such as The Inquirer post numerous articles daily. Sure, there is an editorial process there, but they have a large volume of material coming their way. Not all of the material posted is of the same quality. On a slow day, they post junk anyway.

If people keep visiting a site and see the same stuff all the time, they lose interest.

Aside from the above concerns, I have another. I am not certain that I want to see 'filtered'/modified versions of articles. The ones that I posted fairly represented views that I thought had merit and were in keeping with the venue. Perhaps I was mistaken, but the articles received thoughtful responses. They struck a note with some people.

Committee approved material has a tendency to be bland and often has the life sucked right out of it. More than just the direct censorship applied by the committee, there sets in a process whereby authors begin to self-censor so that they have a greater chance of passing review.

Given the nature of the community here, I think that a filtering process might very well reduce the quality (and certainly the quantity) of articles posted here.

Observations, posted 7 Jan 2004 at 02:53 UTC by mslicker » (Journeyer)

The charter does not seem so bad.

I don't understand the purpose of the distinctions "Master", "Journeyer", "Apprentice". Was this intended as a mentoring program? I've never seen it actually play out here. Is it a matter of status, encouraging someone to try to "climb the ladder" so to speak?

Dualities:

  • Elitism produces anti-elitism.
  • Censorship produces anti-censorship.
There is nothing wrong a group choosing its members. However many of the proposed solutions may be simply causing friction instead making any actual improvements in quality of the content.

It is easy to spam an article discussion, as my demonstration shows. Threaded discussion may be an improvement here, perhaps with a moderation scheme. Perhaps not a rating system, but simply a label "personal attack", "troll", ect.

MSlicker's wise and thoughtful demonstration, posted 7 Jan 2004 at 10:26 UTC by chalst » (Master)

Ah, so your spamming of the discussion thread of one of the most valuable contributions so far made to the thinking on where advogato should go next was a piece of enlightenment, making advogatans aware of important truths they were otherwise ignorant of?

Apology accepted for your other piece of wise and thoughtful input, btw.

Threaded articles, posted 7 Jan 2004 at 10:36 UTC by chalst » (Master)

mslicker suggests that threaded discussions avoid the risk of spaming. Joel Spolsky argues against threaded discussions (I can't find where he says this, but this discussion gives some arguments), and I entirely agree. Do we really want to become slashdot without the readership?

I think if you have a comment you want to make that you think isn't for everyone reading an article thread, the place to do so is in your diary.

Re: Observations, posted 7 Jan 2004 at 13:48 UTC by tk » (Observer)

mslicker:

I don't understand the purpose of the distinctions "Master", "Journeyer", "Apprentice". Was this intended as a mentoring program? I've never seen it actually play out here. Is it a matter of status, encouraging someone to try to "climb the ladder" so to speak?

"Master", "Journeyer" and "Apprentice" are supposed to be approximate indicators of a person's skill, his contributions to open source, etc. They are extremely rough, but raph seems afraid that adding any more indicators will clutter up the interface.

Elitism produces anti-elitism.
Censorship produces anti-censorship.

Duh.

By the way..., posted 8 Jan 2004 at 03:15 UTC by tk » (Observer)

...if everyone can stop seeing everything on earth as part of a big political struggle (as can be seen with use of words like "elitism" and "censorship"), then the world will be a better place.

*sigh*

Re: .... everything above this., posted 9 Jan 2004 at 15:19 UTC by diablod3 » (Apprentice)

I agree that we need a review process, but I would go for something more kuro5hin like, except instead of everyone having an equal vote, Apprentice 'shave a vote that is worth 1 Apprentices, a Journeyer's vote would be worth 2 Apprentices, and a Master's vote would be worth 4 Apprentices. This does have a slight elitism side to it, but since kuro5hin lists everyone who voted what way, it would provide a strong anti-elitism and anti-censorship balancing force as well.

Some one also above said catigories. Maybe we need a kuro5hin-ish catagory system too?

I'd like ratings, as for diaries, posted 9 Jan 2004 at 23:14 UTC by slamb » (Journeyer)

I don't think it's enough to just say "a master posted this article, so it's okay", or even "X masters approved this article". Because some of the least interesting articles can be posted by masters, even people very close to the basis of the trust metric.

Okay, this politically correct, abstract talk when everyone is really thinking of concrete articles is bugging me. So I'm going to mention actual articles posted by actual people. Hope I don't offend anyone too much.

I hated an alternative definition of 'Free SOFTware Developers'. As a physics student, I'm very much interested in talk of fusion energy, but I don't think this is the appropriate place for it. (Try slashdot. More general. Fusion definitely falls under "News for nerds".) Neither is advogato the place for religious ramblings. Or for talk of cookies, delicious though they might be. And I think that articles should be well-written and focused, so even if advogato were an appropriate place for all three, they would still not belong in the same article. It meandered from inappropriate topic to inappropiate topic, and its replies had almost no relation to it at all.

If advogato had the same metric for articles as it does for diary ratings, I think I wouldn't have seen this article. With the other solutions, I still would have.

(Provided, of course, that the ratings are actually updated once in a while. I understand that the diary ones haven't been recently, for whatever reason.)

I think that a rating system would also filter out trolls like Dear Mr Goates, as would virtually any review system.

I'm not sure how articles like Last minute to submit for Codecon 2004 would fare in the rating system. They're not very exciting (it's the papers people want to read, not that the deadline is approaching), but they are important for some people to see (or those papers won't be there), and this is the appropriate forum for them. It would be a shame if they were modded too low.

Discussions like GNU-Info Bad, Man(1) Good would probably fare well in this system. Good. That article is interesting and important for people to see and consider when creating documentation.

Basic newbie help questions like rpm --version -> RPM Version 4.1.1 would be modded into oblivion. Good.

Articles like A Relational Model of Programming are harder to peg. He may have been describing a useful concept, but that was not really demonstrated. As is, it might be interesting to a very small subset of the audience (very theoretical-focused computer science people) and boring as hell to the majority (who are more interested in applied CS / software engineering). I'm not sure about the specifics of the rating system, but if it uses the trust metric like I think it does, it would probably do well here, too. Those theoretical people would be "close" to each other as far as the metric is concerned, and thus they would "correctly" see it rated much higher than the rest of us.

your 'hatred' that someone else touches upon your beloved subject , posted 10 Jan 2004 at 02:39 UTC by badvogato » (Master)

...is quite understandable. But only that this writer has graduated from engineering physics a decade ago. And he only knows now, in English, that plasma from the 'plasma physics' wasn't much older than his own namesake. That was a most interesting revelation bestowed to him by the grace of globalization. He has no desire, whatsoever, to treat Physics as seriously as you or others from Slashdot are entitled to. Still, if you have more than passing interest in "fusion energy" research, i wouldn't waste much time on slashdotting if i were you...

Father, Son and Holy ghost, posted 10 Jan 2004 at 02:56 UTC by sye » (Journeyer)

What anti-macassars, lentils and Dr. Edward Anti-Teller gotta do with Fusion? You tell me...

badvogato..., posted 10 Jan 2004 at 07:27 UTC by slamb » (Journeyer)

I didn't like your article, yes. But please don't read into that too much about how seriously I do or do not treat physics (my beloved subject?), any doubt of your credentials , or a lack of respect for you, as a developer or otherwise.

This is what I was afraid of when mentioning specific articles, but this whole discussion is just silly otherwise - I hate dancing around what you're talking about. You can't design an article review system unless you know what you're trying to accomplish, and I think that means mentioning examples.

How can the quality of articles be improved?, posted 10 Jan 2004 at 13:40 UTC by kai » (Journeyer)

I was initially drawn to advogato out of interest for raphs work on libart and the articles he posted regularly. While I still enjoy reading the diary entries at Advogato from time to time, my interest for the articles which are posted here has almost vanished. The writing style is often of exceptionally low quality and while I don't expect masterpieces like the essays by Paul Graham I would like to see more articles which are still worth reading a month from now, and I don't think the situation needs to stay as bad as it is now.

First of all I have to admit that I don't understand the details of the trustmetrics system hence I can't say whether the improvements outlined in lkcl's article would work out but I actually doubt the problem needs such a sophisticated technical solution at all.

I beleive that almost all of those who are submitting articles don't want to annoy the advogato audience with dull and boring articles which are lengthy, hard to understand by any but a selected few and full of spelling mistakes. Most people would gladly accept criticism which leads to better articles and a more positive response from readers and ultimately helps them improve their overall writing skills.

Of course there are some obvious trolls and posts which should have been submitted as diary entries (some only posted as articles by accident by people relatively new to advogato). As slamb points out about any review process should weed those out easily.

I like chalst's idea of implementing some kind of "refereeing process" which allows people to have their articles reviewed before posting to the front page. In my opinion this process doesn't have to be mandatory or even overly technical in nature. I for one would surely like to have my articles read through by some people with good writing skills (in particular native speakers) and receive some comments on how to make them more effective - if there is any help at all - before throwing them at the crowd.

What advogato is also lacking are guidelines for potential article writers: How to write a good article? How long should the teaser paragraph be and how to communiate the essence of an article in an effective way? What topics are the people who read advogato actually interested in? Are there some essays to look at which are generally regarded as excellent?

So far I haven't written a single article for advogato myself mainly because I simply don't want to spend my time writing something (especially in a language foreign to me) which nobody is actually interested in. I cetainly don't want to bore the kernel hackers or dbms coders around here with my musings on user interfaces, user testing and object oriented programming (or whatever). A way to submit article outlines or ideas before writing them would surely help _me_.

Regarding announcements and calls for papers it's not quite clear whether these are wanted on advogato at all. If so, it would make a lot of sense to have a seperate section for these or just show their headline, not the teaser paragraph. Maybe there should also be a place for newbie questions or recommendations wanted which could be of interest to a wider audience but don't belong in an article ?

If technical modification to advogato are out of question I think using infomal category "tags" could help, like "[Advogato] Time for Article Review Process?" or "[Events] CodeCon registration started today", "[FAQ] Use vi or (X)Emacs ?" and establising a simple peer review mailing list for submitting articles to.

I'd also like to see people not engage in personal flamewars in article responses but maybe that's just me =°) (Not that it's bothers me too much though)

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!

X
Share this page