Hi (:

SOCKPUPPET ALERT!

@BloodPhilia Hi (:

only time will tell @TomWij, Google circles uses Twitter, Google Reader and your contact list, surely that has a great overlap with Facebook

@Nyuszika7H Sockpuppet? :P

besides unless content is only to be found in Facebook's walled garden, Google probably has already indexed it

@BloodPhilia you're rchern's sockpuppet q:

@Nyuszika7H I don't get it =/

and as Chris Dixon from Hunch.com told in a presentation at Google, it only takes about 20 questions to figure out pretty well what kind of type of person you are

@BloodPhilia rchern (a.k.a. Rebecca Chernoff) uses backwards smilies and oy. (:

now if Google get's your permission to connect some dots to recommend you content, surely they don't need everything from Facebook to do so

@Nyuszika7H Oh hahaha XD

oy!

Oy with the poodles already!

Hello, everyone.

@IvoFlipse: His wording for B is right, although traffic != participation != what drives SE...

Well, from the community/site/q&a building perspective, money from advertisements is a different section which is outside my interest. :D

I should've read more than just looking at the chart of bit.ly/hV61rj.

Hmm, didn't see the reply pane yet at the right side of Twitter, still, it's annoying that everyone receives the tweets that are part of your conversation... :D

@IvoFlipse I really like things based on recommendations (when they work).

I don't know what Hunch actually does

but after answering some questions, the topics they thought I'd like were spot on

Bruce Willis movies, Webcomics, MMORPGs (well WoW), Browsers, Touchscreen smartphones, Online email services (Gmail)

I could write a whole blog spot just about recommendation systems, what I like and what bothers me.

Eg. Google Reader has no "not interested" on things you are subscribed on (they assume you are interested in every sing news article), Google News has no buttons at all and just shows:

> Recommendations are based on what you have searched for and clicked on in the past, and they will improve over time as you use Google News.

Didn't seem to kick in yet after some months... :-/

I use My6Sense on my phone to find 'random' stuff

and try Google Reader's recommended stuff

I've explained my thoughts on that and Cadmus (Personal Twitter Trending Topics) before, it's in the log if you care to read about it, that's why I want to write my own well thought out recommendations system. I've got the tools, it just requires me to really dedicate to spending time on it...

Yeah, Google Reader's recommended stuff works well

Jinni is nice for movies, but I haven't watched that site and moves much lately though. I should spent more time on it, or better said, I should watch more movies...

Grooveshark's radio might work well, haven't listed that much to it, again, show listen more music, or better said, learn to put up my music after I start my computer...

And I think that's it, perhaps there is some other recommendation thing that I forget about, like Facebook Friend Search, but those became trivial and less important.

@IvoFlipse: I've also spent some time at Hunch.

Pandora's radio is awesome, but off course it's blocked now

LinkedIn knows very well who you're likely to want to invite

#1 Places to find your soulmate, because I've seen some other romantic articles. :P

#2 - #5 all tech related (browsers, console video games, laptops, cell phones), I dislike the shopping recommendations there though...

Perhaps, I'm still at university but can see how that site might be useful later, I tend to think it's more for America though...

#6 Favorite colors, hmm, I already have those since I was young...

#7 Ubuntu, I'm not into Linux.

#8 Hairstyles, err...

#9 - #11 Cameras (looked into article), mp3 players (have one, perhaps because I looked at headphone recommendations?), iPhone Games (uhm, no...)

I tend to think Hunch works well for most of it, apart from #6, #8 and #11.

I think the issue is that the content is bad there...

Or there should at least be a filter for shopping recommendations. :D

Obviously you want to only see shopping recommendations when you actually want to shop for some thing. Or fun things when you want to have some fun. Or ...

they just use the 20 questions that most likely cover everything they need to know

@IvoFlipse: No, they can do it on a lot of questions, it works with classifiers: bit.ly/f8zQMV

as soon as I want to do some data crunching on my data, I'll know where to find you ;)

@IvoFlipse: Maybe, I don't follow the Data Mining course at University though as the exam tends to be hard from what I've heard. But my Bachelor project is pretty related: bugprediction.wordpress.com

I'm happy to have chosen that one, it prepares me for writing my News recommendation system.

interesting :)

I might make the clinic I do the dog project with contact UA to get help from students in the form of a project

though I might as well do it myself off course

Next blog post coming up this week, I've not worked on it for some time because of a project I had to finish before the exams, the exams themself, the hospital, catching up with courses and meeting my project tutor again. Need to catch up now by doing more work but it will be okay...

I should follow a course on data mining or statistics, but I'd rather spent a month locked up with a book, though I fear that's not going to cut it

About the News recommendation system, I've had these conversations in the transcript: chat.stackexchange.com/transcript/message/614773#614773 chat.stackexchange.com/transcript/message/171475#171475 chat.stackexchange.com/transcript/message/163041#163041

Very short feature list: Follow stuff. Vote topics/words up/down. Select interval in time. Get recommendations/hot things for a topic or a combination of topics.

but how do you recommend new stuff, how do you discover new things?

Very short example: I want to know the most recent about shocks in Japan but not about the reactors: [last hour] + japan + shocks - reactor - fukushima; then I can sort based on popularity/relevance/time/...

@IvoFlipse Having a lot of sources, classify them.

yes, but how would you get access to these sources?

TDT?

@ThomasMcDonald: If I understand correctly, that's how Google News combines multiple news articles about the same thing under one header?

video.l3.fbcdn.net/cfs-l3-snc6/81489/34/…

off course, you're not google :P

@ThomasMcDonald: tf-idf dropped in our conversation about classifying my bug severities, I learned that it's more costly.

Still, I'm more after the thing I explained in my feature list and examples, TDT and/or TF-IDF would result in less articles because they are combined...

last I checked @ThomasMcDonald Netflix is doing pretty good ;)

besides, newspapers often have articles you don't want to read

but the quality of the posts that you like is what keeps you buying it

@ThomasMcDonald Well, Jinni (off-spin from Netflix) works for me. I do have heard about the article you are trying to refer too (can't find it, it was about spikes in their recommendation)...

Hello Napoleon Dynamite!

@IvoFlipse Yeah, I don't want to be Google. I just want to transform the information overflow of things I'm interested in in something more handy...

@IvoFlipse: There you have it! :-)

the thing is @TomWij you should optimize for actual users, not for the OCD types we are, that read more feeds/articles than we have actual time to do that :P

@IvoFlipse: Yeah, actual users would benefit too.

Read the three links about previous conversations, especially that point out why I dislike Cadmus.

Anyhow, I know my system wouldn't be perfect at first, but it will at least me more useful...

That's based on the 1000+ people that I follow on Twitter.

SXSW + Japan are things going on right now, I'm interested in SXSW and it's nice to follow the stuff about Japan to be able to talk about it as it's simply one of the most hottest topics at the moment.

But then... iPad, Twitter and the later Facebook... UGH These are there every day, I wish I could just click a - button to make these occur less, or click a button to figure out the people that cause these topics to trend and see what other topics they trend about.

I simply don't follow a 1000+ people

if you post too much posts and most of them are crap, I boot you :P

Quake and Tsunami, perhaps I could learn my system to group these under Japan if there are enough tweets that tell the system that they belong together at this moment of time.

I'm with ya there @ivo 1000+ followings is too much for me.

I do have too many feeds in Google Reader though @KronoS, so I'm considering booting Gizmodo+Engadget+Slashgear

And then there's SXSWi which I should be able to merge to SXSW, and I dunno what #austin is but if I could just click on it and get the trending topics within #austin I would quickly figure it out.

I don't use google reader. I'm busy enough with allthe other stuff I do at the moment.

@IvoFlipse I use my Twitter the way I use RSS, I don't like using Twitter the same way as Facebook.

Facebook = Private & Friends, Twitter = Public & News & Tech & ...

@IvoFlipse The thing is, booting them could make me miss information.

Surely the sites you mentioned (might) cause Twitter, iPad and Facebook to trend so high, still, they do have nice tweets which are less likely to be written by other sources.

They have for me.

well I just booted them + Ars Technica and Techcrunch

I'm sure people will share it or I'll read it indirectly thanks to all the network effects

@ThomasMcDonald if I get a dollar for every time I read the same piece of news, I'd be rich ^^

I didn't actually read 100k articles, but I skimmed over 100k headlines since then

@ThomasMcDonald Sure, and that's why Twitter, iPad and Facebook could still be trending when I remove them. So I boot again, and again, and again, and again...

And suddenly I'm losing the interesting news.

I'm trying to miss more news, so I can finally get things done

So I rather follow more, filter and thus see more interesting news. As cutting down on interesting news isn't reaching the effect what I'm using the sites for...

instead of reading 100k articles, I could have written blog posts and more importantly: read at least 2 books more and programmed more

that's 70 days rights?

@IvoFlipse I don't like reading books, I need content for my blog posts and I could have programmed my system to only read the interesting stuff and not care about the crap or spent time on booting a lot of interesting stuff.

1600 posts/day

I think @ThomasMcDonald it comes from me rapidly looping through suggested feeds, looking for interesting posts. If I spot enough articles I like, I subscribe, but I want to get a decent sample

how did you count that from 2008 @TomWij?

@IvoFlipse: I guess it's registration date.

Strange yours starts from 1 Jan.

exactly

@ThomasMcDonald Actually some more, but I guess I've done those from the websites itself. Those were to my dad, computer news about hardware relevant to his computer store.

Ah, OCD. :D

Yeah, (s) would be better, but then repetition OCD kicks in.

The proper way would be something like (mail_count > 1 ? 's':'')

I've worked in a Dilbert comic for 2 years, I had the time to read GR

It's just... I can understand that my system is only handy to people who like to follow a lot of content.

I actually see it more likely to be used by a news company than normal users.

For normal users I would have to build up a database of feeds, but I don't have the money to support that overflow of information. Nor would the system work well for about any feed...

That last paragraph was about my system.

@TomWij like I said: if a certain percentage of what you recommend is superb and new, users will come back

@IvoFlipse Sure, but I'm not Google, Google News or Google Reader. I can't support such a database...

and sadly, there's things like Flipboard and such that already do this

Perhaps limit it to Tech as most likely those would be my users.

well the thing is, if you can limit yourself to more popular articles, then the chance of having something interesting is higher

the downside of that is that the chances are high someone already read it through their own channels

@IvoFlipse: But does it allow timeline, trending topics, trending topics within trending topics, filtering, recommendations and all that stuff I explained in my 3 links and above?

manual filtering is far too much hazzle

that's why you need the social graph

You can't filter based on your friends.

no but if news pops up in 3 of yours friends, chances are you'll like that as well

@ChrisF, for some reason I can't launch your app anymore :(

No, because some of my friends have iPads, others like Football, some use Linux, ...

@IvoFlipse Oh. It's not updated so I can't have broken anything ;)

it's not going to be the only determinant @ThomasMcDonald

It is a single instance application - check that you've not still got a copy running. Though it should bring it back if you have.

I'll reinstall

@IvoFlipse Well, you somehow have to tell the system you don't like X before it will show/recommend X less.

In short: I want to maximize the awareness of things I'm interested in and minimize the information overflow (without decreasing the amount of interesting things I receive), I haven't find a system that allows me to do this till this moment.

you partially already do that if you read some articles but don't read others

even when I reinstall it, it doesn't work @ChrisF

which indicates one of it's dependencies must be crapped out

@IvoFlipse Can you define "doesn't work"? Have you told it where to find your music?

@IvoFlipse Well, that doesn't seem to work well...

#5th item in Google Reader:

iOS 4.3 jailbreak on Windows now possible with Sn0wbreeze 2.3

@ThomasMcDonald if you view everyone you follow on Twitter as a friend, you're likely to like something retweeted by friends

it crashes as soon as I open it, some Windows error thing hangs over it :\

See how in that article header I dislike iOS, I might like jailbreaking and I surely like Windows.

man somethings really f-ed @ChrisF :(

@IvoFlipse That's serious.

Can you take a screen shot?

well I haven't had random crashes (after pulling out all USB devices), but can't come out of hibernate anymore :S

So, I would downvote the iOS topic (I want less), upvote the jailbreaking (I want more [well, in reality I don't, I've got nothing to break, just example]) and leave Windows as is as I already have a sufficient share of Windows and wouldn't want less or more.

The only thing it depends on that I don't have any control over is Windows Media Player.

think it might be due to virtual c# too :\

if you look at what it installed

@IvoFlipse: Can you check the event log for a crash dump location?

I don't use any facebook based recommendation services @ThomasMcDonald ;) but that's mostly due to no 'tech' friends on Facebook

Try starting with Process Monitor (to check for access denied issues) and Dependency Walker (to check for DLLs it failed to load) if there's no dump...

DRIVER_POWER_STATE_FAILURE

how would I find that if the problem doesn't occur?

or do you mean when launching Chris's program?

@IvoFlipse: When launching his application. Ah, you said something about an hardware issue too, I seemed to have skipped that thinking about various things...

besides, Google Reader uses some of my Gmail contacts too

which include Stack Exchange users

@IvoFlipse: Still, recommendations based on friends wouldn't work well in my case, it would show what my friends are interested in, not what I am interested in.

It's just, we can keep arguing about it but it boils down to the point that I'll have to write it (perhaps after I've done university I might have more time) and see how it works out...

Example from the last 5 minutes:

scientificamerican.com/…

(I've cheated by using topsy.com/s/Japan?order=date&window=h; which again implements another small set of features I described)

So much sites, all with a small set of different features, to do the same thing but not how I would want it... I'm an unsatifsied customer. :P

@IvoFlipse What is that event from?

It could be related to the "allow the computer to spare energy for this device" setting in device manager.

I've tried changing that but didn't really help :S

and I can't figure out what product was really causing the problem and more importantly why

Also, if you want to enumerate them, use powercfg -devicequery followed by the parameter of the list you want to check:

  -DEVICEQUERY
              Return a list of devices that meet the specified criteria.

              Usage: POWERCFG -DEVICEQUERY <queryflags>

              <queryflags>  Secifies one of the following criteria:

              wake_from_S1_supported  Return all devices that support waking the
                                      system from a light sleep state.
              wake_from_S2_supported  Return all devices that support waking the
                                      system from a deeper sleep state.

@IvoFlipse I'm not sure, but the dump might show that.

msdn.microsoft.com/en-us/library/ff559329(v=vs.85).aspx

btw how do I filter that process monitor?

perhaps we should have a SU question about using that properly

it lists far too many things for me to spot the error

Right click provides filter options, useful for filtering unrelated processes away (success/buffer overflow/no more entries/reparse/...)

Perhaps a video guide?

Or hmm, a post is easier to look through though.

if we create youtube video's, Jeff will also turn on support for it on SU

@IvoFlipse - got to go. If you can get more information on the failure let me know - you can get my e-mail from my SU profile.

will do @ChrisF

but I probably won't solve it any time soon :(

@IvoFlipse I could perhaps help you over connect.microsoft.com/site94?

Or anything else that provides such functionality, although that works the best.

@IvoFlipse: The parameters of the bug check DRIVER_POWER_STATE_FAILURE you had refer to the device, so by analyzing the dump file you can figure out which device causes it. It would give a ID that you have to search in regedit and then read the name...

I don't know how to analyze the dump :\

@IvoFlipse: Either send it to me, do it online or install WinDBG, configure it's symbol settings and perform !analyze -v and read through it.

I think the dump is huge :\

How huge?

should I change the settings to make it that small version by default?

Maybe, have you tried compressing?

There might be a lot of zeroed kernel memory that isn't used in the kernel dump.

> Introduced in Windows 7, the nt!TRIAGE_9F_POWER structure provides additional bug check information that might help you determine the cause of this bug check. The structure can provide a list of all outstanding power IRPs, a list of all power IRP worker threads, and a pointer to the delayed system worker queue.

> dt nt!TRIAGE_9F_POWER 82b5dae0 (second parameter of bug check)

was still 100 MB

And then checking !poaction and !thread might reveal the driver name.

but I'll see what happens next time it crashes

Yeah, it's indeed a bit too much, not worth it to upload the dump.

I wonder though if a mini dump would contain the structure and thread information I just said.

Or perhaps the driver is already shown in the call stack revealed by !analyze -v, but I highly doubt it as that is the kernel doing stuff...

well I found it elsewhere, that was pointing at a Logitech ddriver, but not sure whether it was the webcam or the headset

and why my headset would give problems now, because it worked fine for over a month

Check what drivers and applications you installed right before it started crashing.

too much

and I don't know how to get rid of those specific drivers :\

Although WhoCrashed is fairly simple and doesn't go into detailed analysis, you might want to check if it directly points at a driver.

Disable devices, uninstall devices or upgrade drivers.

off course I won't know if Windows will install the right driver the next time :\

or why it's conflicting

Well, if you're lucky someone else also experienced it and they've fixed it, or you just had an older version, or...

But well, we can't tell for sure. Dump analysis is the way to really know...

I blame SP1 :P

night :)

night @Nyuszika7H :)

@IvoFlipse Or any prior update, SP1 itself just added some specific features that most people don't use like RemoteFX and so on...

SP1 hasn't changed anything to the driver architecture either, perhaps a prior update did slightly to support something.

Or a driver broke because it depended on some RTM-specific value.

@IvoFlipse: As for process monitor, I filter out the annoying processes in the left (svchost.exe, chrome.exe, procmon.exe, ...) and filter out the status messages that are irrelevant to the problem (succes, buffer overflow, reparse, no more entries)

Then it shouldn't flood hard anymore. I run the program and then I start filtering to only get the events related to the program and then I trim down.

And start from the bottom up, as then you start from the moment it crashed.

I'll think about writing tutorial posts for the various troubleshooting software programs. Still it bores me like with the data recovery post I was writing.

It is nearly finished now but took longer than expected...

And although it was community wiki, only @Wil helped me in a good way. :-(

you can also write a blog post instead ;)

@TomWij whatwhat...

The mechanical part of the disk recovery question.

Ahh

@IvoFlipse I don't have access to the blog, and writing it in tutorial form would not be interesting. I would have to solve an actual interesting problem...

Like seen here: blogs.technet.com/b/markrussinovich

you could write it like: "How to use Process Monitor like a real Super User!"

and I can easily give you an account

Im annoyed at the moment... Learning MVC 3, can't find any really good resources (guess too new), and the guide is half about MVC, half about accessing the data... and I am not sure what belongs to which (e.g. EF/data, I could use in other apps)... And, from what I have seen so far, I know it is progress, but I feel like a dinosaur with change... I don't like the way it is going!

can a couple of you users vote me to 5K? if nothing else, as a milestone :P

@Wil: Hmm, I think you might be reading the wrong stuff. What is your goal?

Data is part of the M in MVC.

I'm off for now

see y'all tomorrow!

@IvoFlipse: Cya!

@IvoFlipse bye!

@TomWij To learn MVC!... I am years out of date and only really have a few months history of (real) .net programming. When I last did it, I used to create the DB from scratch and have full control over SQL statements. I just feel like I am not comfortable and don't trust EF... it is amazing what I have seen... but... I can't explain

Basically... I now think I can relate to what my Granddad feels like when I try to teach him to use a computer!

@Wil: As for data: You either work the new way with Entities (either made your model by Code, by Database or by some third thing I forgot) or you work the old ways with SQL, DataRows/Columns/Tables and that kind of stuff...

@Wil Not trust it in what way?

SQL and Arrays of Data back then is now LINQ and Entities, and then under the hood Entities map to SQL queries or what else you would like.

... I need to learn (something) but the official guide doesn't really show you where MVC itself starts/ends and EF starts... I mean, one moment it says to use Nuget to install EF code first thingy (which confuses me in itself), the next thing it says to write something which is a linq statement...

Or you could just work with SQL and Arrays of Data and ignore all the new EF stuff, it's all up to you, there just there to make the job easier for you.

Not having to type out the whole query again to select an item.

But just something like db.Articles.Find(123) to get the article with ID 123.

I think I would feel much better just a single guide for bog standard basic MVC, and another using EF (possibly on a console app) and then, when I know the boundries, learn to put it together... (if that makes sense)... right now, I don't know what is MVC and what is EF.

@Wil I think you have a wrong view over MVC