Konrad Förstner - Comments

An OS, a browser, a DNS, my data in the clouds, a protocol: I do like their products + innovation but I'm more and more cautious w/ google

December 14 from Twitter - Comment - Like - Share

Konrad Förstner liked this

Currently even two operating systems - Android and Chrome OS - which might become on at some point. - Konrad Förstner

Lars Juhl Jensen

Windshield splatter analysis with the Galaxy metagenomic pipeline - http://scholar.google.com/scholar...

November 23 from Google Reader - Comment - Like - Share

Michael Kuhn, Neil Saunders, Jan Aerts and 2 other people liked this

Pullquote: "How many species inhabit our immediate surroundings? A straightforward collection technique suitable for answering this question is known to anyone who has ever driven a car at highway speeds. The windshield of a moving vehicle is subjected to numerous insect strikes and can be used as a collection device for representative sampling." - Lars Juhl Jensen

4 more comments

Pullquote: "Although our specimen collection strategy is straightforward, we set ourselves the nontrivial task of taxonomic identification of collected species." - Lars Juhl Jensen

Pullquote: "The list included unexpected entries such as the genus Homo even though the two trips were uneventful." - Lars Juhl Jensen

I am still not sure if I would consider this as a very creative and elegant way of collecting a rich and diverse sample or as the most useless habitat to analyze chosen so far. - Konrad Förstner

Konrad, I agree with you! However, I am sure that this is some of the most subtle, dry, and well-written humor I have read in recent time :-) - Lars Juhl Jensen

Yes, absolutely. :) - Konrad Förstner

Egon Willighagen

Academia vs. Business - http://xkcd.com/664/

November 18 from Google Reader - Comment - Like - Share

Piotr Byzia, Allyson Lister, Konrad Förstner and 6 other people liked this

Call me cynical, but I think they've got it the wrong way around. - Neil Saunders

3 more comments

Neil +1 - Pierre Lindenbaum

appreciation of good code in academia? Neil +1 - Konrad Förstner

To be fair, I think he's talking about it from a CS perspective, rather than a bioinformatics one. - Chris Miller

You know, there are funny parallels with drug discovery biology here, as well. More often, we're seeing new technologies come from startups (which come from an academic background) and not from established programs at larger companies. Is there enough of a market for short, clever bits of code that you could support a startup model in bioinformatics? - Mr. Gunn

Michael Kuhn to Michael's feed, The Life Scientists

Fwd: Sites using the StackOverflow engine http://meta.stackexchange.com/questio... includes http://asksci.com http://sciencestack.com http://neuroknowledge.net (via http://friendfeed.com/yokofak...)

November 3 - Comment - Like - Share

Daniel Mietchen, Matthew Todd, Frank and 3 other people liked this

Thanks to Pierre for finding this. Math Overflow (http://mathoverflow.net/) actually seems to work and has a community, but the listed sites have almost no audience (yet). Seems Ask Sci is slightly ahead, though. - Michael Kuhn

7 more comments

Here are some recent threads from this room that _could_ also work on Ask Sci (et al.): http://friendfeed.com/the-lif... http://friendfeed.com/the-lif... http://friendfeed.com/the-lif... - Michael Kuhn

re Ask Sci: the homepage listed for the the admin is some kind of shopping portal, which is a bit strange - Michael Kuhn

http://sciencestack.com and http://mathoverflow have LaTex formulas support which is a must for a Science-oriented Q/A site. - Jane Breezler

@Jane: some of us in this room do think that StackOverflow-for-scientists would be a good idea, but please adhere to the guidelines of this room: "Try not to spam and if you are associated with a service (and most don't know that), please say so." – If you're associated w/ sciencestack, we'd love to have a discussion about how such a thing could actually work (or why MathOverflow seems to work, and asksci / sciencestack no [yet]) - Michael Kuhn

Isn't "science" too broad for this? Don't we need smaller units (i.e. chemoverflow, biostack?) Lots of the science questions are very general-sciencey. - Matthew Todd

Despite the fact that sometimes good solution come from people in a different field I would agree with Matthew. More specific sides would reduce the noise for the majority of users. - Konrad Förstner

Well, StackOverflow also has a huge range of topics. The problem with e.g. splitting chemistry and biology is that you then don't know where to put your biochemistry questions. ;) The tagging system (interesting / ignored tags) should help with that. - Michael Kuhn

@Michael: Thanks for the notice. + for mentioning the tagging system! @All: Please direct your discussion and suggestions on Stack Overflow for scientists to http://friendfeed.com/science... - Jane Breezler

Matt Wood

Found a bunch of Wave invites ironically hidden in the murky depths of my Wave inbox (hit me up if you want in) #doubleinboxfail

October 14 from Twitter - Comment - Like - Share

Björn Brembs and Pierre Lindenbaum liked this

Love to give wave a try if you have any invitations left. - Nils Reinton

7 more comments

Me too if there are any available. - Konrad Förstner

Love to try it if you have any left! - Simon Coles

Matt, I'd love one too if possible, cheers - Dave Lunt

I hope "a bunch" is enough to send one my way... - Oliver Schuster

I want an invite! Me, me, me! Pick me! - Nina Jansen

if you still have some: andrew.sid.lang AT gmail DOT com ? - Jean-Claude Bradley

fgibson at the gmail dot com - Frank

Matt, there is a Wave "Perhaps we could start a list of people here to be nominated for invites and note those who already have them?" where there are more FFers who would very much like an invite, and you can see who gave invites to whom... firstname.lastname is my account, in case you cannot find it with the search functionality... - Egon Willighagen

Duncan Hull

Citation-Tracker.com - http://citation-tracker.com/account...

October 1 from delicious - Comment - Like - Share

aarontay, Cesar Sanchez, Endre Sebestyen and 4 other people liked this

The goal of this website is to allow researcher to keep track of citations and other mentions of their work on the web. Sites like Google Scholar are invaluable in tracking citations but lack a few important functions: bullet The ability to organize up reported citations: For many academic reports it is important to differentiate citations coming from peer-viewed publications from the citations that come from other sources (e.g. class homeworks, technical reports etc.). bullet The ability to add citations not detected by the online tool. bullet The ability to monitor the web in general for websites that mention a publication: For example, seminar pages, blog posts, and so on. Especially for such ephemeral web pages, Citation Tracker allows researchers to save copies of these pages for future reference. - Duncan Hull

13 more comments

@Duncan - have you tried this out yet? - Allyson Lister

the login procedure is pretty broken as far as I can tell - Simon Cockell

Looking forward to the reviews... e.g. how social is the service? Can one easily share cite counts, etc? - Egon Willighagen

@Allyson not tried it yet (waiting for registration email) but looks like it might be interesting - Duncan Hull

Hi all, I am the creator of the website. I see that you have problems with logging in. Can you provide some details (OS, Browser) so that we can test and fix this? - Panos Ipeirotis

This looks interesting, I'm going to give it a test ride! - Björn Brembs

@Panos I click on "sign up" fill in the details and receive no email - Duncan Hull

@Panos I could sign up fine, but couldn't login (Firefox 3.5, OS X 10.6.1) - Safari works though. - Simon Cockell

@Panos: +1 to what @Simon says - I'm either logged on, and cannot perform any next steps (because the page does not change), or I'm not logged on. Firefox 3.5.3 and IE 8 on Vista. Will try other OS and browsers soon. - Allyson Lister

I'm logged in, but pretty amazed I can't add publications by PMID or similar?!? Also Google Scholar searches only return a max of 100 entries, this seems to be a limitation if I appear at position 101 in the search? - Daniel Swan

@Panos Sign up and log in worked fine. One bug so far: I got a white page after importing selected publications from Google Scholar. But the import worked - when I manually navigated back to the dashboard the publications were displayed. (Firefox 3.0.15 ... yes, a proper working version of 3.5 takes some time for OpenBSD) - Konrad Förstner

I imported my publications from Google Scholar and everything seemed to work. Not quite sure, yet, how do disambiguate duplicates... - Björn Brembs

Thanks for the feedback! We will be debugging the login issue. This was pretty much a tool for personal use, which I decided to open it up to others, so there are definitely bugs that I could not find through personal use. - Panos Ipeirotis

For feature requests, I would appreciate if you could enter them using the feedback system, so that we can put them in the queue and start implementing them. The tool was focused on Computer scientists and business school scholars, so functionality that is considered necessary for other fields (e.g., pubmed id for biomed researchers) was not necessary. - Panos Ipeirotis

Michael Kuhn

The Ultimate Productivity Blog - http://productiveblog.tumblr.com/

September 23 from Bookmarklet - Comment - Like - Share

Andreas Matern, Allyson Lister, Pierre Lindenbaum and 3 other people liked this

I was just thinking "that sounds like a time-waster", but they were one step ahead :-) - Neil Saunders

Great, that really helped me!!! - Konrad Förstner

Deepak Singh

Invite codes for Infochimps.org - http://mndoci.com/2009...

September 22 from business|bytes|genes|molecules - Comment - Like - Share

Mr. Gunn, Daniel Swan, Bill Anderson and 8 other people liked this

First 50 people to use the invite code in the blog post can sign up for an Infochimps account - Deepak Singh

9 more comments

Done, thanx. - Egon Willighagen

Me too, thanks Deepak - Pierre Lindenbaum

Thanks. - Jan Aerts

Thanks! - Endre Sebestyen

and me, thanks - Neil Saunders

Cool, many thanks! - Konrad Förstner

Thanks Deepak! - Rob Syme

thank you, deepak, :) - chaz2b

Anytime ... thank Flip Kromer actually. I am just the conduit :) - Deepak Singh

Thanks Deepak, having a look around right now... I never knew there was that much data about major league baseball available. It puts our obsession with cricket to shame ;) - Daniel Swan

Lars Juhl Jensen

Dear lazyweb. Which framework would you use to build a web site for a fairly large institute with several groups and why? It would need to be able to do access control to have both public and private pages, allow specific people to edit certain pages, and be able to have pages that are automatically generated from a database.

July 23 - Comment - Like - Share

Claudia Koltzenburg, Chris Lasher, Daniel Rowley and 8 other people liked this

sounds like a case for Plone: http://plone.org/ It's a CMS built on top of Zope, so it's highly extensible. But the list of CMS is long... - Michael Kuhn

22 more comments

Thanks Michael :) - Lars Juhl Jensen

We were using Zope and now switched to cherrypy + cheetah: technically zope is python but depending on what you want to do you might find yourself trapped quite rapidly - Yann Abraham

This functionality is what blog3 will have - besides the science 2.0 functionality - if we get funded: http://docs.google.com/Doc... - Björn Brembs

Drupal? - Duncan Hull

it's only a personal opinion, but Plone is a hideous abomination best left alone (this is a quasi-religious prejudice, much like my one against emacs)... Drupal or Joomla will almost certainly do everything you want, with appropriate plugins (Drupal has a very extensive community behind it) - Simon Cockell

I had a long pretty bad experience with the Zope 2.x framework. Too much boilerplate code, too much under-the-hood tinkering. As a result, Plone is extremely cumbersome to maintain. Less experience with Drupal, but a far better one. If you are doing content management only, definitely look at Drupal. I heard good things about Django for deploying web-based applications. - Iddo Friedberg

Your requirements are exactly what current CMS do. I agree with Simon, dont go with plone. Drupal - http://drupal.org/ or MODx http://modxcms.com/ - Frank

I can vouch against anything like Plone/Zope. Go with Drupal or Joomla. - Paulo Nuin

A US government favorite for such purposes is SharePoint </ducks> - Brian Haugen

Sharepoint is horrendous. I have to use it every day. Don't ever mention that word again :) - Frank from iPhone

Sharepoint is the worse software from MSFT I ever used. It's worse than Me, Vista, Outlook, and any other crap they produce. Unfortunately the company I work for uses it. - Paulo Nuin

Plone is horrific. Drupal is good. I used to like Joomla a lot - but watch out for 3rd party extensions, they often have terrible security issues. A lot of wikis have CMS-worthy features too. I like DokuWiki, but TikiWiki is very full-featured - http://info.tikiwiki.org/tiki-in.... Compare and contrast at CMS Matrix (http://cmsmatrix.org/) or WikiMatrix (http://www.wikimatrix.org/). - Neil Saunders

Thanks everyone! - Lars Juhl Jensen

I agree with the majority that Drupal is very good, but it would be even better if it used a scripting language for customization rather than a web-based GUI. There are way, way, way too many checkboxes that you have to remember to tick and, as a result, you have to be something of a Drupal expert to get the customization right. It is a bit like one of those attempts to write a GUI for a command line program where the programmer just puts a checkbox for every single command line option. - Matt Leifer

Well, I for one definitely recommend that you *do* evaluate Plone. Your security requirements are exactly the sort of use-case that the Zope platform was designed to handle. We've had a role-based security model as a core component for about a decade now. If you want to see who else is using Plone, check this out: http://plone.net/sites - Michael R. Bernstein

Also, this report comparing different CMSs might be useful: http://www.idealware.org/compari... - Michael R. Bernstein

Drupal - Charles Tanton

Hi Lars. If you have money to spend, it seems like these guys have made nice CMS stuff for the University of Copenhagen before: http://www.magenta-aps.dk (mostly Plone and some custom CMS) - Anders Norgaard

Konrad Förstner did something with Plone (I think for the Online Symposium we had at EMBL some time ago) -- you could also ask him for his experience - Michael Kuhn

Right, I used Plone for the Online EMBL PhD symposium (a static version can be found here http://phdsymposium.embl.org/onlines...). This is already some years ago (so before Plone 3) and there were quite some improvements/changes since that. For us it worked out quite fine - the only issues IMO is that its resource demanding and it was a little bit slow on our old OpenBSD server... more... - Konrad Förstner

I'd use Drupal for this, http://drupal.org. It has a vibrant community of contributors, many of them professionals associated with high profile companies. Drupal is GPL with the exception of a few themes and of course custom modules. Access control/permissions are as fine grained as you want them to be, you can set up workflows and content types to accommodate what you state, and then... more... - Christoph Weber

+1 for Plone. It has great users/groups functionality and it's very easy to use. - Nate Aune

+1 Joomla, we are using it in our institute web site. - Khader Shameer

Michael Hartl

Brauche demnächst einen Schlafplatz für zwei Personen in Basel. Wer kennt nette Menschen mit Ein bisschen Platz? #fb

August 20 from Twitter - Comment - Like - Share

Wenn Du offen für so etwas bist versuche mal http://www.couchsurfing.org - Konrad Förstner

Neil Saunders

So, there are other lifestream applications which aggregate feeds and allow comments. Profilactic, Storytlr, to name two. Are there others? Are they any good? What would it take for you to migrate to them? Or are we all off to Facebook?

August 10 - Comment - Like - Share

chrisofspades, M F, Morton Fox and 29 other people liked this

Which ones are OpenSource? If the FriendFeed software was OpenSource, then that would be a certain guarantee for future freedom, as people would set up alternatives in no time, forcing any provider to play friendly with the users... (not implying FF is not doing that!) - Egon Willighagen

35 more comments

are we being a bit quick off the mark? do we know for sure that FF is going away? And maybe it will but take awhile and something else will come along in the mean time. Unless I have missed news that says otherwise I am not rushing into anything. - suelibrarian

It's conceivable, but unlikely, that FF will be maintained. Even just leaving the servers running takes money, staff and time. And very similar services have come (and in some cases, gone) already. I'm not rushing out of here either, but the time will come to look at other options. - Neil Saunders

I think it's safer to assume right now that FF is not going to continue in it's current form. I agree with Neil, other options need to be explored. I don't want to have to sign into Facebook (an increasingly rare event) to keep in touch with everyone here. - Daniel Swan

God no. I had nicely partitioned Facebook into being mostly for family and RL friends. Adding the "friends" I have here would just confuse the conversations such as they are. The main issues I have with FB is that the "groups" are so closed. As soon as the news from the groups are enabled into the news streams they will become more active. I mostly inhabit FB to find out what my 17yr old is up to. Increasingly she isnt on there either. - suelibrarian

It seems unlikely that FF would be maintained - particularly just for us. I'm with Sue, its less the privacy issues that bother me as the mixing of conversations that benefit from being (gently) partitioned. I don't know whether FF might consider open sourcing an earlier version of their code but I don't think anything else really has the functionality. What we do have is the community and critical mass to make it work if we can find the right place... - Cameron Neylon

Noserub is the main Open Source lifestreaming app that I know of. It is not as feature rich (e.g. no groups at the moment) nor as fast as FF, but you can follow someone's FF from a noserub installation as well as following people from other noserub installations. See http://noserub.com and http://identoo.com for an example of a public installation. - Matt Leifer

Noserub looks potentially interesting as an open source platform. Would be interested to know what people with web management expertise think of it as a tool? - Cameron Neylon

Facebook is a big NO for me. Thanks for telling alternatives, I'll surely try them all... - Marcos Marado from fftogo

The gov. agency I work for has been developing a FF-like application for a while now. See my profile at http://me.edu.au/p/nlothian. We do aim to open source it, but that's probably a way off at the moment. But if anyone is looking at trying to develop this from scratch, DM me, because I'd love to try and make something happen. - Nick Lothian

One problem with noserub is that it is ridiculously slow at the moment, especially if you are used to FriendFeed levels of performance. However, it is the closest to FF of any of the open source social media projects that I know of. Also, the federated approach is definitely something I would like to see in a future FF replacement. - Matt Leifer

Streamy. - Meg v. Meg v. 1.0.0.1

Streamy certainly has a better demo video ;) - JSNFLMNG

Hmm. Streamy looks nice but I just tested it for 10 mins, around 9 of which were spent staring at an unresponsive browser, waiting for my CPU fan to stop howling. - Neil Saunders

I was playing with http://noserub.com / http://identoo.com some month ago - I think this will be the place to be. Elgg.org might be also an alternative that I will test soon. - Konrad Förstner

I wonder if elgg is close enough that the community could make it what we want? http://elgg.org/about.php - Richard Akerman

I have installed both Elgg and noserub. Elgg is not really very close to FF because it is trying to be an all-purpose social networking engine. It is more flexible than noserub but would take quite a lot of coding to turn it into a FF clone. - Matt Leifer

Maybe AmpliFeeder? http://amplifeeder.com/index... Haven't tried it, and I think it is missing the comments part. - Endre Sebestyen

@Matt - maybe some plugins could fill the gap http://elgg.org/plugins.php - Konrad Förstner

Agree that noserub seemed unresponsive, didn't try streamy. But presumably enterprise levels of responsiveness takes enterprise level investment? Please prove me wrong, I'd be delighted. - Cameron Neylon

Are you talking about http://identoo.com? Works okay for me. The general advantage of the noserub protocol is the fact that it is decentralized so it's scalable http://noserub.com/quick-f... - Konrad Förstner

Does anyone know how many regular users FF have? - Ola

I don't think facebook will work for the kind of info sharing we have here. It would be too much noise for existing facebook friends, which tend to be more personal friends and less professional acquaintances, and the personal stuff would be too much noise for our science-focused group here. As most of you know, I am a big fan of the federated approach, but has it improved enough in usability to make it feasible if FF shut down tomorrow? - Mr. Gunn

I'd be willing to put cash into a pool to pay someone with serious web scaling credibility in this community (Matt Wood? Neil Saunders? there are people out there, I'm just not really sure who is best qualified) to assess the scalability and suitability of Noserub/Elgg or any other appropriate framework to actually work for the general research community and to cost development pathways. I'd really like a professional opinion on the options. - Cameron Neylon

I'd say Matt is your man for web scaling credibility/development from our community. Today, I find myself hopeful that FF might just be maintained "as is" and we can stop worrying, since Paul writes "I don't want to see it disappear". - Neil Saunders

I'm hoping that too, Neil, but it just seems like too much of a "buy and bury" move to me. - Mr. Gunn

I will be interested to see what happens. I've enjoyed FF, but perhaps I'm showing my age by finding the excitement over it (and the anxiety over its removal) to be a little overblown. As a Usenet veteran, I find the lack of threaded comments rather primitive; the feed aggregation is interesting and the Ajax is great, but I'd be surprised if they aren't replicated elsewhere. IMO, worries that the community will dissipate are excessive: communities rise and fall all the time. - Ian Holmes

Despite my many posts to the contrary, I'm not too concerned either :-) I like FF, I think it's a great demo of how the right technology can enable conversation within/between communities. But it's only a website. There are multiple ways to find people and information - FF is just a particularly effective one. - Neil Saunders

For me identoo is winning: it lacks "hide" to be a usable replacement, and then lot's of features that will come with time. Being opensource, we can easilly affect its roadmap even if not with code ;-) - Marcos Marado from fftogo

I believe that the FF team honestly wants to keep FF around, but it no longer matters what they want -- and what FaceBook doesn't want is another site that makes them look bad, even if they own it. FF is doomed, mark my words, doomed. Dooooomed. I'm with Cameron, I'd happily contribute to a fund to pay for a professional opinion on how to build and pay for a science-centric FF clone. If... more... - Bill Hooker

what about Google Wave? Will it be replacement of FF? - Alexey

ReadWriteWeb article on Streamy: http://www.readwriteweb.com/archive... - Walter Jessen

Nick Lothian - I'd be interesting in talking with you about what neds to be built. Cameron Neylon - I know a few things about web scaling and I would be happy to help evaluate frameworks. - Jordan M

Tons of options out there people http://lifestreamblog.com/create... Covering Lifetreaming services, scripts, and apps has been my passion for over two years. You got questions? I should have answers. - Mark Krynsky

what? no mention of Lifestream.fm?. shocking. - Avatar X

You could just add Lists to Facebook just like you do in here and add FF people into that list. - Manuel Mas

Liking that identoo at least has all the same services. - Mr. Gunn

mridul

As Classrooms Go Digital, Textbooks May Become History - http://www.nytimes.com/2009...

August 9 from Bookmarklet - Comment - Like - Share

Jean-Claude Bradley, Meryn Stol and Daniel Mietchen liked this

"a Beyond Textbooks initiative encourages teachers to create — and share — lessons that incorporate their own PowerPoint presentations, along with videos and research materials they find by sifting through reliable Internet sites." - mridul from Bookmarklet

Another resource for this kind of material is http://wikibooks.org/ and http://wikiversity.org/ - Konrad Förstner

Pierre Lindenbaum

Firefox 3.5. For the first time my geo-location was asked : http://www.openstreetbrowser.org

August 7 from Twitter - Comment - Like - Share

Mr. Gunn, Konrad Förstner, Egon Willighagen and 2 other people liked this

Still like to know how I can contribute to #openstreetmap without a #gps device - Egon Willighagen

You can simple add items that are missing (post boxes, pharmacies etc.) using the client. For this you usually don't need the exact position but can estimate it. - Konrad Förstner

Michael Nielsen

If you wanted to scan a thousand books, how would you do it? I'm looking into digitizing my library.

August 2 - Comment - Like - Share

Christopher Granade, Fiacre O'Duinn, Brian Ahier and 5 other people liked this

Maybe look up how Google is doing it? - Meryn Stol

45 more comments

If you bought any of those books from amazon, then some of them may be searchable through your "amazon library". Having said that, it's spotty coverage (in terms of which books are indexed). Same thing for google. - Ilya Grigorik

I would first look on certain websites (which I would obviously never use myself and certainly don't know the URLs of) to check whether somebody has already scanned them. - Matt Leifer

Meryn - Done that. Not much help... - Michael Nielsen

Ilya - I probably bought a couple hundred (or more) from Amazon. SO that's a great idea. - Michael Nielsen

Incidentally, I found http://sealedabstract.com/... very helpful on this subject. But it'd still be hundreds of hours of work, minimum. - Michael Nielsen

I never realized I could search through my Amazon purchases. Interesting. Quite a value-add from Amazon! - Meryn Stol

Looks like Google will give you full access if you scan the barcodes: http://booksearch.blogspot.com/2009... Not clear if this is useful for re-reading, or if it's possible to download... - Michael Nielsen

I tried that.. no dice. It only works for books that have been "open sourced" by publisher or are not under copyright, which is to say, basically nothing. Ok, not entirely true, but of limited value.. - Ilya Grigorik

Thanks, Ilya. That's one weird Google blog post, then. ("It works for all the books except the ones it doesn't work for. That is, most of them.") - Michael Nielsen

the way I understood the Google barcode scan was that since it appears you own a physical copy of the book, you can have access to a digital copy. Am I wrong in this, because I was going to do it for my books. - Kimber Scott from BuddyFeed

Kimber, yes, but the amount of digitized books on Google Books is limited. Actually, the bar code does not constitute proof of ownership. You can do fulltext search through all books on Google Books. But with "my library" you can limit your search to a group of books. (Pedro, yeah exactly) - Meryn Stol

I just gave it a try. I scanned a couple of barcodes but its really just a faster way to create a list. I don't think the access rights changes at all. - Pedro Beltrao

That's the thing.. The scan just gives you the ISBN, which you can find anywhere anyway - it doesn't prove that you own the book in any way, shape or form. For this reason, only books that are not copyrighted are available for search. - Ilya Grigorik

Ilya, there are copyrighted books searchable on Google Book search. Just not all of them because not all publishers have signed up for the program, or have not committed their whole library. Full search does not mean you can read it in full though! - Meryn Stol

Ah, well that's good news! Now we just need to get all the other publishers on the same bandwagon. ;-) - Ilya Grigorik

Major project! I guess ideally, there'd be a way to prove that you own a legitimate copy of a book which would then grant you access to a digital version. Something like scan and upload one page. - Neil Saunders

Neil - Yeah, that'd be very nice. Sadly, I haven't found anything like that. - Michael Nielsen

yeah, but that's not going to happen either - publishers are more than happy to charge you twice for the same material - once for print and once for e. Related, Springer now has a $25 print on demand service available to patrons of libraries that buy the big ebook packages - we only have it for medical books right now, or i'd try it - Christina Pikas

Maybe a pragmatic 80/20 split - re-buy the most important books (those that are not from Amazon) as e-books; then decide if you want to cut off the binding on the remaining books and have em scanned (as suggested in article), or make do with a semi good scan without damaging the book (pay a student to run them through a copy/scanning machine). - Christof TD

Please try to save the integrity of the book. You may find that they will be quite valuable. Were some of them gifts? Do they have inscriptions you want to keep? Some things to consider. - Melanie Reed

On 2nd thought you'd better wait a little (machines like the Google book scanner should become more readily available soon) - reading about the Google Book Scanning Machine: http://www.npr.org/blogs... - Christof TD

Here we go - Book Drive Pro (looks quite industrial; but still doesn't solve the page-turning problem) - http://pro.atiz.com/ or http://mini.atiz.com/ (video on the pro version site shows the scanning workflow); ~several hundred hours for 1000 books... at least it gives you a high-quality outcome without damaging books - Christof TD

Hmm - you could build this thing yourself (Atiz Book Drive Pro): make a V-shape cradle, including a plastic V-shape that comes down on the book (like in the video), several mounts for two DSLR cameras and a little black tent with lights -> then run an edge-detection algorithm on the pictures taken using the Python Imaging Library; this leaves only finding an open-source OCR library. - Christof TD

Christof - For all but a tiny fraction of the books (which I'll probably keep), I'm not worried about destroying them as part of the scanning process. Your 80/20 suggestion is a really interesting one... - Michael Nielsen

You could donate the destroyed books to a local school science library; maybe the library will find the resources to glue them back together again :-) - Christof TD

More likely the library will waste valuable staff time and resources evaluating and then discarding a book the school cannot use. Please don't use libraries as a dumping ground! Thank you. - D0r0th34

If you are a DIY person check out http://www.geocities.jp/takasci... and http://www.instructables.com/id... . My personal solution (I got of most my books some years ago - some I kept and scanned them manually) is a change of mindset. Most books are just dead trees that catch dust and are need much to rarely to excuse... more... - Konrad Förstner

Thanks for the link: http://www.instructables.com/id... ...this was exactly what I was thinking (see video at that link for workflow); 20 min. for each book is not too bad. Update: their open source software takes 3 hours to process a book into a PDF. - Christof TD

Here's my stupid idea: 1) The most difficult part: Build a little stand that automatically flips through the pages - probably doesn't need to be as sophisticated as in the Instructables link above. 2) Take a HD video cam and make a video of the pages being flipped. 3) Write a little script that takes screencaps from the video every other second, after the pages have been flipped. 4) Send all the screencaps to Evernote to use their OCR and search. - Victor / Mendeley Team

But then again, it would probably be easier to just sync a photo camera with the page flipping :-) Like I said, stupid idea! - Victor / Mendeley Team

Here are the heavyweights: Treventus Book Scanner Robot (at Bavarian State Library, Munich) - http://www.treventus.com/booksca... and Google/ Microsoft supposedly use Kirtas high-performance scanners - http://content.zdnet.com/2346-95... ... ah, the page-flipper uses suction: http://www.kirtas.com/ - Christof TD

Video of Treventus ScanRobot in action: http://www.youtube.com/watch... - Christof TD

Ah, sweet.. Now i just need one of those for my house! :-) - Ilya Grigorik

buy a high speed scanner - abuzarhamza

abuzarhamza - Even at 20 pages per minute, with roughly 300,000 pages that's roughly 300 hours of scanning... - Michael Nielsen

How about this: (1) create a new organization called the Nielsen University library (2) offer your library's collection to google to become part of their scanned collection (3) make the one condition of the offer to google be that you get the scans for private use - Jonathan Eisen

I suppose that this would not be a good time to point out to you that ownership of the physical volume doesn't entitle you to scan the content and create a digital version; that's what got Google into trouble in the first place. You, as an individual, haven't the right to create (even for personal, private use) digital versions of content where other entities hold the rights. - Jill O'Neill

Don't forget that Michael is covered by Canadian law, which is significantly more lenient on copyright than US law. That's why I suggested finding pirate copies on the net. I believe that under Canadian law might not be liable for downloading them, but only for uploading. However, the situation is complicated so don't quote me on that. After downloading, you are in no better or worse shape than if you had scanned the books yourself. - Matt Leifer

I note - I had my tongue planted deeply in my cheek when I said he should get google to do the scanning - just in case people cannot see where my tongue was - Jonathan Eisen

Jonathan: Your original comment gave me, at least, quite a smile :-). Sadly, I can't imagine Google would find much in my library that they don't already have. Maybe the classic "Venus on the Half Shell"... - Michael Nielsen

Jill - Is it true in Canada that you can't legally transfer between media in this way? I know I've been told that in some jurisdictions this definitely _is_ allowed, but I don't recall in which jurisdictions. - Michael Nielsen

Michael: I was speaking from a US-centric view. That said, I would be surprised if -- to pick a publisher at random -- HarperCollins (whether US or Canada) would look favorably on such scanning activity. Libraries are permitted to scan for purposes of preservation under specific conditions, but I am not sure that the same permission has ever legislatively been extended to private... more... - Jill O'Neill

Have you read Victor Vinge's Rainbows End? ok it's a bit extreme but it works for gene sequencing :) - Alexei

the fujitsu snapscan is reasonably good but v manual and destructive, I've taken to guillotining spines and feeding pages while watching tv .. ~ couple hrs / book. End result not that good, big bulky pdfs. Getting eprints would be much better. - Alexei

Thankyou, by the way, to everyone who's been commenting on this! - Michael Nielsen

If you want to use all free and open source tools, read this blog post: http://philikon.wordpress.com/2009... - Nate Aune

Pierre Lindenbaum

Wikipedia offline - http://www.rob-matthews.com/index...

June 6 from Bookmarklet - Comment - Like - Share

Peter Murray, Berci Mesko, MD, novoseek and 8 other people liked this

Not even close to the right size, is it? - Mr. Gunn

3 more comments

That's just the featured articles... http://en.wikipedia.org/wiki... - Andrew Su

Featured articles are about 0.08% of the total articles, so you'd need something like a thousand of those "books" to print the whole of WP. - Bill Hooker

I do not want to know what happens to the book when an article is updated :D - novoseek

I just stumbled across this visualization of Wikipedia's size: http://en.wikipedia.org/wiki... - Konrad Förstner

The Life Scientists: Wladimir Labeikovsky

The pleasure and importance of printed journals : Nature - http://www.nature.com/nature...

July 1 from Bookmarklet - Comment - Like - Share

Bora Zivkovic, Jason Tsai, Mary Canady and 17 other people liked this

Someone needs to upgrade from dial-up...... And get off their ridiculously high horse. - Wladimir Labeikovsky from Bookmarklet

29 more comments

and learn about print on demand. - Pedro Beltrao

Liking the comments rather than the letter, obviously :-) I guess they're entitled to their (ridiculous, wrong-headed, irrelevant) opinion. - Neil Saunders

The only thing I can identify with is this act of discovering while browsing. There is no very good equivalent online for this (yet) - Pedro Beltrao

The quality comment is what kills me. As if somehow cellulose and ink enhance the quality of the peer review... - Wladimir Labeikovsky

All his points are killing me: You want to browse during breakfast? Has the guy ever heard of a "browser"? Been reading my news et al. during breakfast for the last 10 years. Laptops FTW! And yes, Wladimir: how does cellulose and ink make a journal any better, I wonder? I think we should get together and write a pointed reply to Nature! - Björn Brembs

Wow, Bjorn, feel free to include my name on the letter, but I wonder if it's worth the bother. It's not like we'd even have any common ground to discuss. - Mr. Gunn

@Mr. Gunn: Worth smorth. Consider the fun! :-) - Björn Brembs

It would be fun! - Mr. Gunn

Only problem may be that the letter which is the most fun writing won't be the one getting published, lol :-) - Björn Brembs

I've had a letter published in the print edition. I know to what degree those things are edited before they hit the page. - Mr. Gunn

'Until Nature comes up with a new business model, to read this story in full you will need to login or make a payment' - Andrew Spong

@Mr. Gunn: So you'd think "upgrade from dial-up" and "ridiculously high horse" would be edited out? :-) That cracked me up :-) - Björn Brembs

Here's my first draft: http://etherpad.com/FZ2N9FKECI - Björn Brembs

As for random browsing, wikis have it by default, and it's the outcome of many a search on a journal's site or on other parts of the web. - Daniel Mietchen

Despite the fact that printed journals contribute only a small fraction of the total amount of paper I think it is a question of mind set and awareness. Why get hundreds of printed pages of a journal if you only skim through a small part of it and read only a tiny part at all? That's quite a waste of resources (paper, ink, energy for production and transportation). Yes, paper... more... - Konrad Förstner

One more on random browsing: http://search.twitter.com/search... . - Daniel Mietchen

Yes, Konrad. See also http://ff.im/4GRMM. - Daniel Mietchen

After the fall of Genset in 2006, they started to put their whole library in the garbage. I've tried to save a part of the journals (see this picture http://www.flickr.com/photos... ). I sent some e-mails to the local libraries, to the centers of research, etc... Nobody asked for those journals and at the end, all those books ended into the bin. - Pierre Lindenbaum

@Pierre - Really a shame, maybe try next time bookcrossing.com or freecycle.org at least for the books. I had a similar experience when I wanted to donate some books in quite good shape to the local library. They didn't take them for whatever reason. - Konrad Förstner

That's pretty good, Björn, but I wouldn't even concede the point about discovery. I think discovery through re-sharing of bookmarks (connotea/citeulike/Mendeley(soon)) and through article use stats (Mendeley, PLoS) works better than randomly paging through a magazine. - Mr. Gunn

@Mr. Gunn, I don't know about that. We pick topics and people tend to be somewhat close to what interest us. I can't understand half of what is in a regular issue of Nature but I guess that if I ever browsed it in print I would spend more time on articles that I would never even look at the abstract. I am not even saying that one way is better than the other .. just that my online browsing habits are much more focused than would be if I had to check paper versions. - Pedro Beltrao

Yeah, I'm definitely more focused online, but the serendipity of print limits you to that one article, whereas recommendation online can come from the whole body of work. Maybe we need a good way to browse recommendations? - Mr. Gunn

Is there an analogy about paging through a set of recommendations from trusted people rather than through someone (the editors) with whom you have no direct connection? - Cameron Neylon

I page through F1000 listings, FF bookmarks (citeUlike and connotea) from others, and my broad search terms. Would that count? My subjective impression is that this leads to a broader sample of the literature than if I were just browsing the hip journals. But it certainly also feels less leisurely and comfy :-) - Björn Brembs

Great conversation. I am a strong believer that social bookmarking and networking tools can do a better job of nudging serendipity along than flipping through a paper journal. Use of these tools increases the chances of coming across relevant discoveries in adjacent fields. (shameless plug: http://2collab.com/ is designed with this goal in mind.) That said, my former position at Lulu.com left me a firm believer that print-on-demand is the way to go whenever print is prefered. - Michael Habib

"the proportion of reading by U.S. science faculty from browsing decreased in recent years, replaced by other means of learning about articles that are read. While the proportion of readings decreased over the years, however, that number of readings found by browsing remains about the same: 88 readings in 1977 and 95 in 2005. Readings from searches increased from 17 to 78 readings between these two years." - Tab. 1 in http://www.dlib.org/dlib... . - Daniel Mietchen

"the quality of these prestigious journals could gradually decline to the standard of many of today's web-only journals." What a flippiant remark, as above how does the medium of information transfer effect the quality of the content. What a dinausaur, lets start a charitiable fund to buy them an iphone - although will will have to advertise this via a telegram - Frank

And indeed given some effort they might actually rise to the standard of several of today's web-only journals, one can at least hope... - Cameron Neylon

I'll play devil's advocate here. My first reaction was rather like those above, but there's room for all, no? I liked getting American Scientist in print, though have now opted for online. But there are places laptops aren't very practical. Breakfast isn't one. I don't want my serendipity forced on me, either, cf Pedro. Comment about quality might have to do with the lower cost barrier for a new journal to enter web-only vs print, but didn't like idea that quality for established journals would decline! - Heather

From a practical standpoint I find that when I'm reading something in print, my mindset is different and I have a longer attention span and can actually get through a long article without realizing there's something else I need to look at or do online. I think there is still a place for printed journals, but more for reviews or 'popular' science such as GEN or the Scientist. If it's a primary research paper you really need to read for your work or research, I think the media doesn't matter. - Mary Canady

Michael Nielsen

StackExchange™—The Stack Overflow Knowledge Exchange Platform - http://www.stackexchange.com/

June 29 from delicious - Comment - Like - Share

Morton Fox, Konrad Förstner, Nick Boucart and 12 other people liked this

"The Stack Overflow Knowledge Exchange Platform Designed by Jeff Atwood and Joel Spolsky, Stack Overflow has rapidly become the best place for programmers to get answers to technical questions. Now there's a way to get the same kind of site for your audience. Bird calls? Stamp collecting? It's up to you The idea of a knowledge exchange, running the same software as Stack Overflow, can be applied to just about any subject matter. With StackExchange™, you can run a site with all the same features that made Stack Overflow successful. Customizable look and feel StackExchange allows you to override style sheets, choose color schemes, and provide your own logo. You can use a stackexchange.com URL (for example, philately.stackexchange.com) or your own top-level domain. You can also insert arbitrary HTML into the parts of the page where stackoverflow.com would display advertising." - Michael Nielsen

10 more comments

I've only just come across the Stack Overflow podcast recently.. this has been mentioned a few times since then. I like the way they are opening up not just the platform but the data too... - Daniel Swan

Some useful commentary, including links to similar open source products, at: http://news.ycombinator.com/item... - Michael Nielsen

Wow. So shall we try SO for scientists? - Michael Kuhn

But SO works because it's a site for IT geeks. I don't think it would work for science. What would you ask here ? A bibliographic reference ? Some help with a protocol ?... - Pierre Lindenbaum

I would ask the same kind of questions we are asking and answering now in the life scientists room. - Michael Kuhn

@michael But most of the time, we don't ask any question here :-) - Pierre Lindenbaum

@Pierre -- we have had a few cases though, and people have got really good useful answers. My own example: http://friendfeed.com/billhoo.... Perhaps it's just a matter of FF'ers slowly developing the habit of bringing questions here? - Bill Hooker

It would work for chemistry if there were a quick, intuitive way to share a) data b) pictures and c) video along with posts on SO. If there's any kind of barrier to that, people won't bother. Sharing code is so easy in comparison. I guess also that a chemical lab is quite a social place - lots of people around you you can ask first. Don't programmers often work in physical isolation, where their collaborators are online? - Matthew Todd

I think it is hard to predict if scientist in general would accept it as a tool - if the critical mass of a certain scientific community accepts it it might become a useful resource for them. Another direction: It could be useful for teaching - students ask certain questions that are answered by others. - Konrad Förstner

@Konrad: about 'teaching'. I think this is the goal of Nature Scitable: http://www.nature.com/scitable - Pierre Lindenbaum

@Pierre - thanks for the hint - but I guess there are many different ways of approaching the online teaching aspect like there are many ways of dealing with programming related topics. Maybe I haven't found the SO like functionality in Scitable but what make SO so attractive to me is its simplicity and pure focus on "I have a question - who has the best answer?". - Konrad Förstner

Pawel Szczesny

Ignore Everybody: and 39 Other Keys to Creativity - http://www.amazon.com/dp...

June 18 from Amazon.com wish list - Comment - Like - Share

Richard Akerman liked this

In any case someone wants to read the online version: http://www.gapingvoid.com/Moveabl... - Pawel Szczesny

Nice hint! Hugh MacLeod's cartoons are great (if you can handle cynicism) especially his venn diagrams (one is shown in the online version of the book). - Konrad Förstner

Iddo Friedberg to Iddo's feed, The Life Scientists

Hi all. Please have a look at this almost-completed poster (printing tomorrow). Will be going to OCCBIO2009 and BOSC. Feedback appreciated.

June 9 - Comment - Like - Share

Pierre Lindenbaum, Rob Syme, Pawel Szczesny and 12 other people liked this

I would aim for more symmetry in the boxes: try to have them the same width or at least vertically-aligned. Colour schemes are a matter for individual taste of course :-) My preference would be keep the dark box header + inverse text, make the rest black (or dark blue) on white. At least use consistent box header colour within a column. Also, don't forget some kind of contact information (email/URL/institution). - Neil Saunders

10 more comments

Ditto what Neil said. You've got way too much color - choose two, or three at most. I think the biopython logo gives you a nice color scheme to work with - gold, black, and yellow. Make the logos at left and right 75% smaller, make your title 125% bigger, and don't be afraid of whitespace. Eliminate the gray bar around the author names. I might even consider just doing a black outline for each of the boxes (or no outline at all), with the orange headers - Chris Miller

Never mind the colors: Even if the graphical design was perfect, I'm not sure there's much on this poster that would cause me to even slow down while browsing through a zillion posters. I suggest at least making the biopython URL large enough to be seen from far away; someone might remember and go have a look at the site during a boring talk... - Eric Jain

Another idea: since the poster does not present a sequential sequence of ideas from left to right, you might try a "radial" design: summarise the main points in a central box (or circle), arrange the other features around it on "spokes". - Neil Saunders

Those are all really good ideas. If you can free up some room, I'd try to keep the logos about that size or even make them bigger, otherwise ditto everything above. To get the benefit of Neil's radial idea without too much reformatting, I'd put the Python and BioPython boxes in the center column and the three that are now in the center into the left hand column. I'd also increase the... more... - Bill Hooker

I would propose to use less and lighter colors - maybe play around with this: http://www.colorschemer.com/online... Maybe use simple gradients - that looks usually quite fancy. - Konrad Förstner

Minor issues: Why is there an empty line after "HMM" in the orange box? Maybe use bullet points in this box; Maybe add a white space after the bullet point in the big green box - Konrad Förstner

I can't comment on the content, but for the design I would: 1) use a grid system for the layout; 2) use a colour scheme (http://colorschemedesigner.com/, http://kuler.adobe.com/#themes... or http://www.colourlovers.com/blog...); 3) Use up to a maximum of two different fonts (e.g. one for the titles and one for the main text). It is good to leave space in between the boxes and... more... - Paul Bacchus

Agreed with Eric's comments on the URL; combine it with the Biopython box and increase the size. After all, the purpose of the poster is to get people to visit the site and join the community. For the sequence box, I would show translation instead of transcription, as it's too easy to do replace("T", "U") without a framework. For the SeqIO box, how about example code instead of the table? Fastq to Fasta conversion is very popular recently. - Brad Chapman

By the way, it's great to see brave souls putting stuff out there for "open peer review", and to see polite, constructive feedback. More of this please! - Neil Saunders

Wow, so many helpful people out there. Thanks you all for taking the time to comment! I will not have time to implement all the comments (I also got quite a few from the Biopythoneers!). Probably attenuate the colors to a better scheme (thanks fro the link!), and standardize boxes' width. I wish I had the time to do the spoke thing. @Paul: one suggestion I will NOT follow up upon is to put my photo up: that would scare people... ;) - Iddo Friedberg

+1 Neil, this is a great example of community in action. - Bill Hooker

Pierre Lindenbaum

SVG version of Aspirin. This xml document was transformed into SVG with my XSLT stylesheet pubchem2svg - http://lindenb.tumblr.com/post...

June 4 from Tumblr - Comment - Like - Share

joergkurtwegner, Yann Abraham, Konrad Förstner and 5 other people liked this

I hope this didn't cause a headache ;) - Konrad Förstner

2 more comments

@Konrad, not really, that was quite simple. (Although the pubchem XML format looks weird, but I guess it come from the conversion from ASN1 to XML on the NCBI side) - Pierre Lindenbaum

ah... "Aspirin"... ok, I get it :-)) - Pierre Lindenbaum

hehe :) - Konrad Förstner

Pedro Beltrao

First postdoc paper formally accepted in PLoS Bio. Its about the evolution of phosphoregulation in yeast species. Time to celebrate :)

May 6 from Twitter - Comment - Like - Share

Chris Miller, Peter Binfield, Bora Zivkovic and 28 other people liked this

Congrats Pedro - nice work! - Jason Stajich

15 more comments

whoohoo! look forward to reading paper (and perhaps blog post?) - Neil Saunders

Well done. Buy one on me (seriously -- will pay you back when I next come by SF) - Jonathan Eisen

Congratulations! Until the concept of journal prestige dies out completely (if it ever does), that's the best bio journal in the world. - Bill Hooker

@Jonathan, what, you buying beers for all authors in PLoS Biology? _Definitely_ need to put that in the marketing materials... (I bet the I.F. would go through the roof -- sorry Bill...) - Andrew Su

Well, the plan was that this was specific for Pedro (and what is I think his first PLoS Bio paper but he can confirm ...). But sure, if there are other first time PLoS Bio authors I am more than happy to buy a beer for them. However, after Pedro there will have to be some restrictions - like I have to be there .... - Jonathan Eisen

congrats Pedro-great work - Abhishek Tiwari

congrads, looking forward to reading - chaz2b

hey, well done! congrats!!! - Thomas Lemberger

Thanks everyone, specially Jonathan for the beer ;) - Pedro Beltrao

congrats ! - Pierre Lindenbaum

Congratulations Pedro! Can you send me a reprint? Oh wait, it's Open Access :-D - Lars Juhl Jensen

Lars FTW. :-) :-) - Bill Hooker

Congrats Pedro! - Duncan Hull

Excellent! - Konrad Förstner

Nice. - Peter Binfield

Congratulations! - Ruchira S. Datta

Egon Willighagen

Slashdot | Wolfram|Alpha's Surprising Terms of Service - http://tech.slashdot.org/article...

May 20 from delicious - Comment - Like - Share

Matthew Todd and Allyson Lister liked this

We shall have no more WolframAlpha mistake screenshots: "Wolfram|Alpha's terms of use are completely different in that it is not a search engine; it's a computational service. The legalese says that they claim copyright on the each results page and require attribution. So for you academics out there, be careful." - Egon Willighagen

Man, this is what I call a "broken license"! - Konrad Förstner

You mean they've copyrighted the phrase "Wolfram|Alpha isn't sure what to do with your input"? - Matthew Todd

Deepak Singh

Some early Wolfram|Alpha driven thoughts - http://mndoci.com/blog...

May 16 from business|bytes|genes|molecules - Comment - Like - Share

Michael Nielsen, Duncan Hull, Sasha Kovaliov(.com) and 25 other people liked this

My immediate impression for anything chemistry related is: ugh, where are the sources? In your example of "ice" - there are a bunch of data with no source. Clicking the source link at the bottom of the page gives a list of stuff which is not necessarily used for the data quoted, and is not broken down per datum, . For more information, you have to apply! We need people to understand the importance of citing sources, which is why WP can be so good. - Matthew Todd

38 more comments

Matthew, don't disagree. But I don't see this and WP as being equivalent either. Systematic knowledge is, in principle, unambiguous. The biggest challenge for WA will be figuring out that boundary and what constitutes formal knowledge. That's why it works very well for physics and chemistry and mathematics. Those properties of ice don't require citation really since they are unambiguous. Biology on the other hand, just to pick an example, isn't quite as absolute - Deepak Singh

It was Iddo not me who noted as first that life sciences data are pretty incomplete :). See also comments here: http://friendfeed.com/wolfram... - Pawel Szczesny

Pawel thanks ... updating my blog backend right now, but will update when things are up and running - Deepak Singh

Worse of incompleteness is that you cannot see where info is coming from, and not fix it either. With Google you just add a new HTML page which correct information... How does that work with WolframAlpha? - Egon Willighagen

http://www.wolframalpha.com/partici... - Deepak Singh

The challenge they face is figuring out how to make this scale. This is curated data at a scale beyond big. How do you do that? They have to learn to trust some sources - Deepak Singh

Deepak, thanx for the link... should have found that myself. - Egon Willighagen

Very nice article. "Add to that sources like Freebase..." Right. And more generally, bringing the web of linked data into Alpha (or the other way around) makes good sense. - François Dongier

I've noticed something a little disturbing in that it doesn't seem to have an internal check that different things should be different. Do a search on "structure of glucose" and "structure of galactose" and then compare D and L forms etc. There seems to be no internal concept of "different chemical/molecular species have different structures" that you would expect to flag up a problem like this. See http://friendfeed.com/cameron... - Cameron Neylon

Cameron - yuck. That's horrendous. - Matthew Todd

Given it was the first and most obvious thing I could think to test that might cause problems it was a bit of a disappointment. But its just the backing data that is the problem. The response to the query is sensible at least in principle. - Cameron Neylon

Meh. You mentioned two things, and it's put them side by side. It's made no attempt to answer your essential query. WP entry answers it OK. What is the added value from WA? - Matthew Todd

First of all the problem that WA is attempting to solve is non-trivial. It is NOT a search engine and looking through an index to deliver a document. It's optimizing functions from different bits of data and computing an answer for you in real time. So it's not perfect right now, but it will take a while to get there. Our responsibility is to make sure the results are accurate (as... more... - Deepak Singh

It will be very interesting to see how efficiently they can incorporate user's feedback to improve Alpha. - François Dongier

Agreed - Deepak Singh

Deepak, I agree with all of that, but it seems odd to have built in no apparent concept of equivalence or difference of chemical species and no concept of chirality. I agree that problem is non-trivial and I think putting up the two structures is a reasonable response at first order. Shows that it understands the query in some form at least. I am surprised though that there doesn't... more... - Cameron Neylon

That is somewhat strange. I've found a few interesting errors as well (demographics mostly). For WA, thee query itself and the actual computation are different steps. From my understanding there is a natural language front end that translates the query into the Mathematica language and then computes on it. So in this case, it seems to have understood the query correctly, but it might be interpreting just part of the query and that part returns the same answer. Definitely worth pointing it out to them. - Deepak Singh

Deepak - I would not say that chemistry data is commonly unambiguous. For a few textbook examples like water info is probably mainly correct from various sources - but the interesting compounds have properties that probably have been measured only once. Data provenance is important - (though even Wikipedia does not have references for most values in the chem boxes right now) - Jean-Claude Bradley

Chemistry is certainly a domain where the computation approach should really shine (compared to Google Search or Wikipedia). So it would indeed be disastrous (but very surprising) if Wolfram Research couldn't find an easy solution to the glucose-galactose problem. - François Dongier

I think the solution is probably reasonably straightforward - they need to use a more sophisticated internal representation of chemical names and structures. And as Deepak and others have said - way of collaboratively working over and improving the data. All things that we know people with the smarts and the tools to deliver - Cameron Neylon

Agreed. Want me to make some connections (unless someone already has good ones)? - Deepak Singh

Overall, i would say the wolfram|alpha is the step in right direction but the road is full of ambiguous signs. - ashish

Sure, Deepak, and let us know how we can help. - Mr. Gunn

Let's see what I can do. - Deepak Singh

@deepak if you have connections on the Wolfram side that seems like the best place to start - get them talking to Tony/PMR/Egon/BLueObelisk etc. and see where we get from there - Cameron Neylon

Would be nice to get some insight in how they work... looking forward to playing with their API, but too much on my list already... the inconsistencies found by several people here in the chemistry area seem to indicate that the actual computation they talk about is rather limited... I would very much see their view on that. - Egon Willighagen

Also see Nova Spivack's feed (http://friendfeed.com/nova ). - Pawel Szczesny

I'm having a hard time picturing the advantage WA could give us at the moment, re chemistry, for example. Cameron has asked for the difference between A and B. What other kinds of questions ought we to be asking WA, if it were performing as we hope it might? Presumably these are not questions simply about 'what is x or y' but more relational questions such as 'what are all the chemicals... more... - Matthew Todd

Did you check these examples? http://www26.wolframalpha.com/example... http://www11.wolframalpha.com/input... - François Dongier

Underwhelming. These are data I can get elsewhere. I guess it draws some graphs for me, using data from I know not where. It specifies the maximum temperature of a concord nose tip to two decimal places, a gaffe for which I would severely reprimand an undergraduate. - Matthew Todd

"Using data from I know not where": at the bottom of each answer, you should see a link "Source information". - François Dongier

Yes - see comment #1 above. No good. - Matthew Todd

Yes, looks like there's room for some improvement with respect to quoting sources. At this stage, I would be very happy if the answers were just useful :-) Not sure full source details would be very hard to provide. I agree having to send an email is not optimal. - François Dongier

Deepak - As it looks to me that WA will not share data under an open license I would personally not submit data to it. I prefer your idea of using it as an interface to Freebase or similar open data sources. Maybe you meant that by saying "the licensing and business models need to evolve", too. - Konrad Förstner

Here's a useful description of some of the Alpha datasets: http://reference.wolfram.com/mathema... - François Dongier

Konrad, yes. That is what I was alluding to. It's not crawling and indexing anything so how it would leverage other sources remains to be seen - Deepak Singh from IM

The blog is down/suspended? - Eric Jain

argh ... thought I was moving off by now and didn't pay my darn invoice - Deepak Singh from IM

thanks for the heads up - Deepak Singh from IM

Konrad Förstner

EMBL puts data in the public domain via CC0 http://sciencecommons.org/weblog...

May 10 - Comment - Like - Share

Pedro Beltrao, Cameron Neylon, Deepak Singh and Pawel Szczesny liked this

I assume Lars and Michael were the driving forces to do so. - Konrad Förstner

Pierre Lindenbaum

100 Publications Every Graduate Student Should Read - http://compgen.blogspot.com/2009...

May 7 from delicious - Comment - Like - Share

John Dupuis and Neil Swainston liked this

I have been wanting for several years now to make a list of 100 important publications that every one of my graduate students should read before they graduate. I will make this list here and will slowly add to it and edit it over the coming months. I will attempt to organize these by discipline. Please email me your suggestions or comments. - Pierre Lindenbaum

For books, Worldcat.org can give a list of editions and what libraries have them: http://www.worldcat.org/oclc... - John Dupuis

My favorite was already mentioned in the comments of the blog post: "Why Most Published Research Findings Are False", John P. A. Ioannidis, PLoS Medicin, http://is.gd/xyKV - Konrad Förstner

Michael Kuhn

@neilfws <table><tr><td valign="center"> ... </td></tr></table> -- cf. http://giveupandusetables.com/

April 24 from Twitter - Comment - Like - Share

Joe Dunckley, Lars Juhl Jensen, Pierre Lindenbaum and 2 other people liked this

yes, more than one guide points out that when 4 lines of HTML become 20+ lines of CSS, you have to question the benefits :-) - Neil Saunders

2 more comments

screw CSS, i never use it (almost) - Christopher Harris

wasn't CSS supposed to abstract away presentation from HTML? which was supposed to make things easier for the semantic web? which will be here any day now? :) - Richard Akerman

Yup, so damn true. Not sure how many hours I wasted with similar stuff... but still I am a CSS idealist... - Konrad Förstner

Roderic Page

#elsevier #challenge #winner Reflect http://www.elseviergrandchallenge.com/team1...

April 21 from Twitter - Comment - Like - Share

Frank, Duncan Hull, Pawel Szczesny and 10 other people liked this

congratulations to the winners (esp. to Lars & Michael ) ! - Pierre Lindenbaum

4 more comments

yay! happy news before going to bed :) - Michael Kuhn

Thanks for breaking the news, Rod! Now let's get taxonomy tagging into Reflect :-) - Lars Juhl Jensen

Lars, Michael congrats - Deepak Singh

Thanks Pierre & Deepak! - Lars Juhl Jensen

Congratulations! - Konrad Förstner

Richard Akerman

why I don't like FriendFeed beta - http://scilib.typepad.com/science...

April 6 from Science Library Pad - Comment - Like - Share

John Dupuis, Polly Potter, Mr. Gunn and 4 other people liked this

need to be improved - kang

Fully agree, especially the lack of source icons is a step backwards. The keyboard shortcuts are quite lovelessly done - I would use them otherwise. - Konrad Förstner

Michael Kuhn

I successfully defended my PhD this morning with the best mark ("summa cum laude").

March 10 - Comment - Like - Share

Lars Juhl Jensen, Chad Davis, Pierre Lindenbaum and 10 other people liked this

I'm not a PhD yet, though (this is Germany!) -- some formalities still tbd. - Michael Kuhn

10 more comments

Congratulations! - Bill Hooker

Congrats anyway! It's all over bar the shouting. Do you feel special yet? - Chris Cotsapas from Nambu

Gratuliere! - Björn Brembs

congratulations :) - Pedro Beltrao

Congrats - Deepak Singh

well done - well deserved (if brief) rest coming? - Cameron Neylon

thanks for the congratulations. :) I'll take two-weeks break between PhD and Post-doc in June. - Michael Kuhn

Congratulations! That was certainly well deserved :) - Lars Juhl Jensen

Congratulations! - Shirley Wu

Way to go! - Nir London

Super! Congratulations! - Konrad Förstner

Konrad Förstner

Ui, "programmable matter". But how close are they already to it? http://www.youtube.com/watch...

Ui, "programmable matter". But how close are they already to it? http://is.gd/lHk4

March 4 from Twitter - Comment - Like - Share

The concept animations are great but they could have made a better effort of showing what is the current state of the art. I have no clue btw :). - Pedro Beltrao

Exactly, or a comment like "expected to be available in 10 years" - Konrad Förstner

Konrad Förstner › Comments