spaetzblog

Stuff that doesn't matter

Migranten in der Schweiz

Monday, 10 February 2014 | Schweiz, Privat | comments

Gestern hat die Schweiz abgestimmt, 50,3% sprechen sich für eine Begrenzung der Einwanderung aus. Zum wird wieder einmal der Röstigraben deutlich. Als Ausländer möchte man doch eher in den französischsprachigen Kantonen wohnen, die stimmen einheitlich gegen die Aufhebung der Personenfreizügikeit:

http://img.welt.de/img/ausland/crop124685943/0278727095-ci3x2l-w620/DWO-schweiz-abstimmung-aw.jpg

(c)Infografik Die Welt, Tagesanzeiger.ch

"Ohne den hohen Ausländeranteil - er liegt in der Schweiz bei rund 23 Prozent - ginge es seinem Land viel besser, behauptet Christoph Blocher, Chefideologe und Vizepräsident der SVP. "Alle Probleme, die wir haben, sind darauf zurückzuführen", sagte Blocher.

Wenn man sich die Reden der SVPler im Fernsehen anhört, sollte man Ihnen allerdings folgende Statistik vor Augen halten. Wenn Sie so weiterpoltern, wird diese Zahl Ausländer in der Schweiz eventuell auch schneller sinken als ihnen lieb ist:

/uploads/Hotellerie_Schweiz.png

(c)BFS, Taschenstatistik der Schweiz 2012

Ich bin gespannt, wie dieses Geschichte weitergeht. Allerdings ist es leichter es von außen als von innen zu beobachten.

Microsoft buys Nokia

Saturday, 07 September 2013 | Research, research, Microsoft | comments

Nokia's 98,000 employees [1] will probably not be delighted to learn about today's announcement to sell the core of Nokia, its mobile phone business and mapping services for a total of 5.44 billion € to Microsoft.

According to the press release, the transferred units represent ca. 50% of Nokia's revenue. The low price tag (€3.79bn cash plus €1.65bn for patent licenses) is amazing: In 2011, Nokia had a turnover of $50.18 billion (2012: $39.91bn). Compare that price/revenue ratio to those of other recent technology company acquisitions (or IPOs) and you'll see how desperate Nokia has become in a very short time.

The man behind the last three years strategy is Steve Elop, who was hired away from Microsoft to lead the ailing firm back to the upper ranks of mobile phone manufacturers in 2010. He has been called a mole and a trojan horse before (indeed, already at the press conference at his inaugeration), an accusation that he has denied. His immediate strategy of focusing solely on Microsoft phone as the single operating system, was heavily critisized by Ex-Nokia executive Tomi Ahonen among many others (see, e.g. The Elop Effect). Elop is now expected, as the MS press release states, to transfer back to Microsoft.

The series of events is peculiar at very least, I have to say. If it looks like a trojan horse, walks like one and quacks like one, ....

As a result of its failed strategy, Nokia has not been doing well during the last years. In the second quarter of 2013 Nokia has been shipping 7.1m Windows smartphones, which is puny compared to Android smartphone sales of Samsung (73.3m), LG (12.1m), Lenovo (11.4m), Huawei (10.2m), or ZTE (10.2m) during the same time frame. [2]

So what is now left of Nokia? The network unit Nokia Solutions and Networks, formerly known as Nokia Siemens Networks and which was founded in 2007 as a joint venture. However, core network hardware is coming under severe pressure as Chinese manufacturers such as Huawei are gaining ground. There seems to be a mapping division left (but not their apps), as well as a division for Advanced Technologies.

So far, consumers and technology firms have not warmed up to Microsoft on their mobile phones. It will be interesting to watch if they will do so now that Nokia and Microsoft are under one roof. And it will be interesting to see what the rest of what is still Nokia will become in the following years (my guess is that it will be snatched up by a Chinese company?).

The reactions have been interesting, from techcrunch.com predicting Elop's ascendency to the throne of Microsoft, to sad eulogies. Nokia's share price jumped from 2.96 to 4.28 on the day of the announcement.

[1]src: Heise.de
[2]src: IDC Worldwide Mobile Phone Tracker, August 7, 2013

A case of plagiarism at ETH

Wednesday, 07 August 2013 | Research, Plagiarism | comments

Much has been written about the increasing levels of article retractions during the last months. Often, peers, reviewers, and student mentors have been criticized for not doing their job. I believe that we do have a responsibility for bringing these cases to light. However, as long as organizations - afraid of negative publicity - fail to enforce strict penalities, the problems will remain.

My previous employer, ETH, just has withdrawn the degree of a student after it found a case of plagiarism. However, she can re-enroll, reclaim her course credits and rewrite her thesis with a new topic. WOW, what a punishment. It reminds me of the story of the 1980s Budapest railway system, where the penalty for dodging the fare was ... the price of a regular ticket. I have heard similar stories of my previous alma mater, the HSG, although I cannot confirm these from my personal experience.

Detecting plagiarism as a mentor of a thesis, or as the reviewer of a paper is hard, it is really hard work. Why would I invest my time and go through the hassle, if I know that fraud will not have serious consequences?

I have been waiting for the report of an Research misconduct committee dragging out their report for more than a year, while the retraction count of a German professor skyrockets to at least 12 retracted articles with no report and no consequences in sight. The same university only took a month to give an already retired professor a mild slap on the wrist for publishing a student's work as article without any mention of said student. Why would I expose myself accusing peers when nothing is going to happen?

I do not aim to punish people for minor errors ("only those who don't work do not make mistakes"). And I certainly do not appreciate the digging out of 30 year old dissertations just for the sake of detecting plagiarism. But organizations should face uncomfortable realities and implement harsh penalities once serious fraudulent activities are detected and made public.

Booking a travel

Monday, 15 April 2013 | travel, web, Digital Markets | comments

Using modern technology platforms to book a travel, using the modern web 2.0 should be a no-brainer, right? This is what I thought yesterday when starting to book a simple travel from Zürich to Hamburg an back.

So first, do your citizenship duty and check out the railway system (yes, I am ecologically biased). German Bahn has a "Switzerland special" for 39€ one-way, great... But their web form only allows to select a German station as departing place and a return trip starting in Switzerland. There is no way to book Switzerland-Germany-Switzerland. grr. So off to http://sbb.ch and book there: enter all travel details, select discount cards etc, just to be informed that (unspecified) parts of the trip Zürich-Hamburg can not be booked and therefore no booking at all is possible. So back to http://bahn.de, skipping all discounts and simply enter Zürich-Hamburg in the regular web mask. It is great that they offer me to enter my Swiss half-fare card or my German half-fare card, but that is an "exclusive or", I can enter either my Swiss or my german half-fare card for a trip spanning both countries. Fail again! Price: 220€ for the discounted cheap tickets bound to a specific train.

So off to the flights: expedia.de offers a direct 230€ Swiss/Lufthansa flight, so I'll go for that. (1.1h rather than 8h of traveling time is more attractive). Enter details, skip rental car and insurance options, just to be informed that the flight is now 270€ rather than 230€. Suckers. Off to http://lastminute.de, booking the very same flight for 230€.

Conclusion? Digital markets might be true enablers, but their execution is still lacking. Also in-transparent price design can be something rather frustrating for the consumer.

When one door closes, another opens

Thursday, 14 February 2013 | Hamburg, Work, Uni Hamburg | comments

I should have written about this quite some time ago, but never got around to actually do it. The last months have been exciting. I handed in my habilitation at ETH Zurich and it passed end December. Starting February 1st, I can now call myself a "Privatdozent ETH", having been awarded the Venia Legendi (Teaching credentials) in Management Sciences.

Joel West, Joachim Henkel, Juliana Sutanto, and of course Georg von Krogh have steered me through the habilitation process, and I will always be grateful for the volunteer effort that they've put into this.

So far so good. But this also implies that my time at the ETH is coming to an end. I actually attempted to remain in Zurich and applied at both ETH and University of Zurich -- to no avail.

I have been applying at a few Universities in Germany during 2012, and out of 5 applications was invited to present at four Universities (no, I will not publicly shame the one University that did not respond for seven months at all, rejecting me with a 2-liner).

Rufannahme Universität Hamburg (Forschung und Lehre)

©Forschung und Lehre 2013

Prompted by this announcement in Forschung und Lehre in February 2013, I am very, very happy to announce that I successfully negotiated with the University of Hamburg in December 2012, and that I am in the process of getting my chair at the University starting in July 2013. In Hamburg, a city of media publishers and technology startups, I will be the heading the Chair of Management, specifically Digital Markets, at the School of Business, Economics and Social Sciences. More specifically, I will be located in the Department of Socioeconomics, the Fachbereich Sozialökonomie.

This implies a lot of changes, all of them exciting, and most of it to the positive:

For the first time in my life I will have a working contract that is not time-limited at the outset. This means a lot to a man whose hair is getting gray and whose kids are growing up. It also means being my own boss for the first time in my life and being the boss of a few others depending on me. This is going to be a tremendous challenge and one that I look forward to.

Research-wise, I will be even freer to pursue the topics that I am curious about. While my current chair of strategic management and innovation is broad enough to let me follow my research interests, digital markets and digital technologies have always been the core of my curiosity. Georg von Krogh has always given me all the academic freedom that I need. Let us not talk about resource endowment at this point though...

Teaching-wise, this will be quite a change. I will have to do more teaching that I ever had done, and I will have to do it in a system that I have not been part of for 13 years. I have already been struggling with lecture directories that require abstracts to be in (exclusively) German for courses that are going to be taught in English as part of an International Master's programme. Oh well, may you live in interesting times...

Family-wise, it is a relief to be living closer to our families, and to actually have family members living in the same city.

I look forward to collaborate with new colleagues, such as Jetta Frost, Mark Heitmann or Dirk Gilbert, and also to be able to renew collaborations with other colleagues such as Jan Lungberg, Bertil Rolandsson, Linus Dahlander, Christina Raasch, or Oliver Alexy.

Economist semars

Monday, 01 October 2012 | research, ETH | comments

Today, I went to a seminar by, from, and for Economists!

Karin Nyborg (University of Oslo) talked about "Cooperation is Relative: Framing and Income Effects with Public Goods", finding that "rich people" contribute to public goods independent of reward, while poor were self-interest driven. Interesting stuff.

This is relevant as we are also looking at contributions to a public good in Collaborative Open Innovation.

Describing standard Linear Public Good game. Decide how much to keep for you, how much to give to the group. The group income is then multiplied by some factor and split between the group. .. Payoff x_i = e_i - c_i + m(Sum(c_k)1/N Double blind, no one will know who has decided how. (not even lab personnel)

Highly endowed (rich) students get 10€, the poor one 5€. Project suceeds only if contributions reach 120NOK (half of all endowments).

The decision variable of how much to contribute is changed in three treatment groups. It is 1) Absolute, that is, "I decide to give 10NOK to the group", it is 2) Relative, "I give 30% of my endowment to the group", or 3) Payoff ("I keep 70 NOK for myself").

As a result 1) and 2) are basically the same. 84%, or 80% "succeeeded", ie reached the treshold, the payoff group was only 67%. Wow, it matters how you formulate your question.

Rich people gave the same in all cases (1,2,3). Poor people contributed most in the "Absolute" case (40NOK), less "Relative" (30NOK) and least in the payoff case (10NOK or so). Overall, it was a nicely and well done experiment and quite cool, overall. I could well imagine doing something similar.

0:1 for LibreOffice

Monday, 10 September 2012 | opensource, Office | comments

I love LibreOffice. In combination with Zotero, it serves me well writing academic articles. Until.... I have to share documents with Non-LibO-using collaborators. Unfortunately, there are still a few of those out there.

Recently my wife told me, she wants Windows and WinWord back, after she lost a few hours of work due to incompatibilities.

I feel that LibreOffice does not get its priorities in the right order. Sure, code cleanup and all that is good. But LibreOffice seems to be mostly thinking in terms of (new) features. Blogs show impressive feature comparison checklists with competitors, such as this one by M. Meeks. The Wiki compares features with MS Office, and each release sports an impressive release note visualizing cool new stuff. But for those poor souls that try to hold out against the MS Office fortress the most important feature is reliable interoperability with .doc and .docx formats. This is tedious and boring work and hardly leads to flashy release notes. Yet, for me -and I would wager many others out there- this is the most important thing, LibO can offer.

It is not a matter of life and death, but students have gotten marks deducted as LibreOffice had introduced spurious superscript in a file when saving as .doc.

About half of the files my wife saves as .doc or .docx refuse to open on her colleagues' MS Office versions (bug 45983). No one seems to know why. All embedded OLE objects, such as charts or spreadsheets are physically lost from the file when saved by LibO as .docx file (bug 51550), an issue that has bitten me bitterly. And when opening a not so complex .doc file, a table was simply missing in Libreoffice (bug 54430 which led my wife to waste quite a few hours, because she first did not notice the lack of the quite crucial table. (and had to track down another version of the file which turned out an obsolete version). To cut a long story short: my wife wants me to install Windows and WinWord, although she prefers LibreOffice's UI. I am sorry, but I can understand her. 0:1 for LibreOffice.

Don't get me wrong, I am grateful for all the work that is being done in and around LibO, and I know about the difficulties of reimplementing a reverse-engineered file format that is handled differently by different MS office versions. The blame lies squarely with Microsoft for not being open here. Yet, I feel that rather than hunting for new features to check of, more effort should go towards those poor souls that actually have to read documents that other people have written and the other way round.

Sollte man das Handelsblattranking boykottieren?

Tuesday, 04 September 2012 | research, ETH, ranking | comments

Momentan ist eine heisse Diskussion im Gange, ob man das Handelsblattranking boykottieren sollte. 291 BWLer haben einen offenen Brief von Margit Osterloh und Alfred Kieser unterzeichnet, der das Ranking kritisiert. Kritisiert wird vor allem die Methodik und die falsche Anreizwirkung auf Wissenschaft und Gesellschaft.

Schwierige Fragen. Zum einen kann und sollte man niemanden zwingen, keine Rankings von öffentlichen Daten zu erstellen. Oder sollte sich diesen entziehen dürfen. Insofern, finde ich ein Boykott eines solchen Rankings nicht gerade befürwortenswert. Sollen Publikationslisten etwa geheim gehalten werden?

Auf der anderen Seite muss die Nützlichkeit solcher Rankings hinterfragt werden. Das HB sagt dass sie ein Tool liefern. Wie es angewendet oder missbraucht wird, ist Sache des Anwenders. Das Argument kennen wir (“guns don’t kill. People do”), aber dass die Existenz solcher Tools einen Einfluss auf das Verhalten Einzelner hat, ist unbestreitbar.

Und ja, ich kenne genug Forscher deren Forschungsagenda sich nach Publizierbarkeit in Journals gerichtet hat. Und bei weitem nicht alle waren junge Foscher ohne lebenslange Anstellung. Ich bemerke dass bei mir selber. Ich kenne interessante Journals, die zu Nischen publizieren die mich interessieren, die eine besonders nutzerfreundliche Open Access Philosophie haben, oder sonst interessant sind. Aber nicht im “SSCI”? Sorry, kein Interesse. Schade? Ja!

Journal Impact factor kommen und gehen. Beispiel Technovation (Elsevier Verlag, Profit Margin 37%): Impact Factor 2011: 3.3, Impact factor 2007: 1.0. Sind Technovationartikel von 2007 über die Zeit qualitativ drei mal besser geworden als noch früher (so wie bei gutem Wein :-) )? Im Ranking ja, denn es betrachtet nur einen einzigen Impactfactor. Oder sind die Editors (und Autoren) nur besser darin geworden ihren IF zu managen? Ohne Self-cites wäre der journal Impact factor statt 3.3 nur 1.7, denn 48% der 424 citations sind self-cites. (Zitationen von Artikel im eigenen Journal, wie von vielen Editoren gefordert, sind ein anderes Thema, das mit der Wichtigkeit von Impact Factors einhergeht). Wo immer es Rankingsysteme und Benchmarks gibt, werden sie “gegamed”.

Im Vergleich, Organization Studies hovert seit 5 Jahren konstant bei 2.0-2.3 IF. Vielleicht hätten sie ihre Autoren einfach bitten müssen mehr eigene Werke zu zitieren? (nur 19% of 277 are self-cites).

Nicht zu vergessen, dass der Impactfactor hier sich nur auf Artikel die in den letzten 2 Jahren publiziert wurden bezieht. Die Publikationsmuster durch lange Review/Publikationsprozesse sind aber extrem unterschiedlich. Ich habe einen Artikel, der 2008 vom Journal akzeptiert und 2010 publiziert wurde. Wie soll so ein Artikel zu einem 2-jährigen Impact Factor beitragen? Mehrmals dauerte es 9-12 Monate bis ich Feedback auf eine Submission erhalten habe. Die zitierten Artikel werden dabei nicht jünger.

Am Beispiel Technovation, die cited Half-Life ist 6.0 years, d.h. mehr als die Hälfte der Citations kommen erst nach 6 Jahren (nur 10.5% der Citatations kamen aus den relevanten letzten 2 Jahren). Org Studies hat eine Citation Halbwertszeit von 8 Jahren und nur 4.5% aller cites kommen aus den letzten 2 Jahren... Andererseits ist mir ein Physical Review Letters von Submission bis Publication in 3 Wochen gelungen. Wie soll man da vergleichen? Äpfel und Birnen gefällig?

Bei jeder Statistikgrundvorlesung die sich mit Powerlaws beschäftigt wird darauf hingewiesen, dass der Mittelwert bei Powerlaws eine denkbar ungeeigneter Wert ist um etwas Vernünftiges auszusagen. Beispiel: Die durschnittliche US Firmengrösse ist 19.0 Angestellte. Super, aber der Modus (häufigste Wert) ist 1, und der Medianwert 3 (Axtell et al 2001)! Was kann ich also aus dem Mittelwert über die meisten einzelnen Firmen aussagen? Und das sieht sehr ähnlich bei Journal Artikeln aus. Was kann ich also aus dem mittelwert eines Power-law-basierten Journal Impact Factors über einen einzelnen Artikel ablesen?

Ich habe genug grottenschlechte Artikel im SMJ gelesen und ebenso viele Perlen in no-name – oder vielmehr besser – no-impact Journals, um zu wissen, dass man vom Journal Impact Factors nicht auf die Qualität der 3-4 individuellen Artikel eines Forschers schliessen kann. Und kommt mir nicht mit double-blind Review. Ich habe genug durch die Augenbinde lunsen können um zu sehen was da passiert (aber das ist ein anderes Thema).

Um es zusammen zu fassen. Werde ich ein Ranking boykottieren? Nein. Finde ich es sinnvoll und unterstützenswert? Nicht unbedingt, und wenn dann nur sehr sehr vorsichtig und begrenzt. Sobald Rankingsysteme existieren, ändert sich das Verhalten der Teilnehmer um sie zu gamen! Insbesondere wenn diese Tools missbraucht werden um Berufungskommissionen, Lehrstuhlevaluationen und Fakultätsbudgets zu erstellen. Insofern vielen Dank an Frau Osterloh und Herrn Kieser! Normalerweise sage ich zynisch, dass solche Rankingverurteilungen immer nur von emeritierten Dinosauriern (Verzeihung :-) ) kommen, die nichts mehr zu verlieren haben. Deshalb freut es mich besonders, in der Boykottliste bekannte Namen von jungen aktiven, und durchaus publizierenden, Mitforschern zu lesen.

P.S. Ich kann es nicht lassen, meinen Lieblingsartikel zu Impact Factors zu verlinken (Hint, Forscher stirbt und möchte in den Himmel kommen. Witzig.) http://genomebiology.com/2008/9/7/107

[UPDATE 7.Sep.12] Das Handelsblattblog berichtet dass die meisten der Boykotteure nicht in den Top-Listen aufgetaucht wären. Natürlich fällt es leichter einen Boykott zu unterzeichnen wenn man -gerechtfertig oder nicht- nicht am meisten davon profitiert. Allein die Tatsache dass 23 von 339 Boykotteuren in den Listen genannt worden wären, zeigt dass nicht nur unproduktive Forscher sich einem Ranking entziehen wollen.

P.P.S. Bei aller Kritik am Ranking, denke immer noch, dass Forscher kein Recht haben sollten sich diesem zu entziehen. Es basiert immerhin auf öffentlichen Daten, selbst wenn die Erhebung mit Hilfe der Forscher geschieht.

Wer innovativ bleiben will, sollte kooperieren

Tuesday, 28 August 2012 | research, publication, ETH | comments

Das ETH Hausjournal ETH Life hat einen kurzen Artikel über unsere Strategic Management Journal Publikation gebracht:

Wer innovativ bleiben will, sollte kooperieren (20 Jul 12)

Schweizer Firmen sind zwar sehr offen für externes Wissen, doch es gibt noch Potential. Vor allem wenn es darum geht, formell mit Partnern zu kooperieren. Denn - richtig dosiert - fördert der «Blick über den Tellerrand» Innovationsgeist und Markterfolg, wie eine Studie von ETH-Forschern zeigt....

Hier ist der ̀ganze <http://www.ethlife.ethz.ch/archive_articles/120720_offene_innovation_ch>`_ Artikel.

Get the Brother MFC-8860DN to scan under Ubuntu 12.04

Friday, 17 August 2012 | Linux, HowTo | comments

I have a Brother MFC-8860DN Printer/Scanner/Fax on my LAN which I tried to get to scan under Ubuntu 12.04. It was quite a journey, but I succeeded in the end. This documents what I have done, as others might suffer the same problem.

  • Download the brscan2 and brscan-skey deb packages for 64 bit. Install.

  • brsaneconfig2 -a name=MFC-8860DN model=MFC-8860DN ip=xx.xx.xx.xx

    (Replace the ip with your static IP address of the scanner. The name argument can be any user friendly term you want, really)

    Theoretically, this is all you are supposed to do. But, alas, no luck yet.

  • Check the output of brsaneconfig2 -q. Does it show your scanner? Good.

  • Check the output of brscan-skey -l. Does it show your scanner as "Active"? Mine showed "Not responded", and it turned out I must have fudged something in the scanner options on the printer web interface earlier that made it non-work. I had to factory-reset my scanner before it worked, ie it showed up here as "Active". This cost me a few hours to find out.

    The most common error here is that people configure the wrong IP address. Do avoid that :).

  • The next hurdle was that it still would not recognize the scanner with xsane, simple-scan, etc..

    sane-find-scanner already reported a possible usb scanner though.

    It turned out that brscan2 installs libraries into /usr/lib64 and they are now supposed to be in /usr/lib/. So, they were not found at all.

    (SANE_DEBUG_DLL=255 scanimage -L can help to find out weird things like this, although it is very verbose)

    Solution: copy or link: /usr/lib64/sane/* to /usr/lib/sane and /usr/lib64/libbr* to /usr/lib/. (credits go to: https://wiki.archlinux.org/index.php/Brother_DCP-7020)

  • I was nearly there. scanimage -L started showing::

    device 'brother2:net1;dev0' is a Brother MFC-8860DN MFC-8860DN

    but trying to actually scan using scanimage, xsane, etc led to::

    open of device brother2:net1 failed: Invalid argument

    It turned out that my copying of the libraries was not done well, as the symlinked files were still pointing to /usr/lib64 locations. Check ls -la /usr/lib/libbr* to see which ones need fixing. Of course, had I left the libraries where dpkg installed then, and simply symlinked them to /usr/lib/ everything would have worked now.

  • I did not have to add entries to /lib/udev/rules.d/40-libsane.rules as you will find in many web sites. a) This file is 60-libsane.rules under current Ubuntu now and b) this is only if you connect it directly via USB. Mine is on the LAN.

  • I did not have to create a /etc/sane.d/brother.conf with an entry starting "usb ...." as you will see in many howtos. Again, this is required for connecting the scanner directly via USB.

Good luck to all you poor souls out there with similar problems. This was horrible to debug, thanks to lacking documentation and obscure error messages. I wish Brother open sourced that crap.