Why despite the EU’s €4.3bn Google Fine, things won’t be fine.

A few days ago, the European Union decided to hit Google with a €4.3bn fine. The reasons put forward by the European Commission focus on the company’s MADA or Mobile Application Distribution Agreement, that all device manufacturers that want to license Google’s apps and include the Google Play Store with their devices are forced to sign.

MADA has been sporadically discussed by the tech community for a few years now, since the 2014 leaks of several MADA agreements from years past. Here is an article at re/code from 2014. Another article from 2014, by The Information, outlines how Google has been modifying its licensing agreements to include more restrictions for OEMs. The EU has been preparing its case against Google for at least 2 years; here is a 2016 article by The Verge discussing the investigation.




The jing-jang of hardware and software support

Ever since the 1980s, a vicious cycle of software and hardware requirements and updates has ‘plagued’ users and maintained high and constant rents to the vendors who systematically collude to render their previous offerings obsolete, while forcing (not enticing) their customers to upgrade to the latest incarnation of their products.

The justification for this form of planned obsolescence has — invariably — involved mention of changing standards, APIs, market and technical requirements, certifications and so on. Perfectly good products have time and again been condemned to retirement only because their manufacturer deemed it uneconomical to maintain support for them.

During the late 1990s and early 2010s, the open source community did much to reverse this process, with varying results. Open source drivers, initially developed by individuals and, later, by small companies aiming to attract revenue from disgruntled users that wanted to preserve their investment in hardware devices, proved that the ‘market’ and ‘economics’ arguments were invalid at best (and deeply hypocritical at worst). Some OEMs quickly embraced open source, providing access to their device driver source code; others maintained a binary blob. Others, when their positioning afforded them the option, maintained their limited support policy.

Sadly the allure of SaaS and cloud-based solutions, the gradual conversion of personal computing equipment to ‘appliances’, both through the proliferation of Android and iOS-based devices, through the integration of erstwhile external components (sound cards, networking, I/O interfaces) and the inevitable demotion of activities previously requiring support for device interconnection (printing is an example), has led to the apparent sidelining of this issue, by both consumers and regulators.

There are many major vendors guilty of this. Canon, who limits support to printers and scanners only a few years after they are released. Native Instruments, who ceases to support hardware devices only a mere 5-6 years post release, citing changing APIs in Windows or macOS, even though those devices are extremely simplistic (the Rig Kontrol 2 and 3 are good examples here) and the community might be able to keep providing support for them for years, had the company released the specifications for writing a device driver. Apple and Microsoft itself, the former removing new features when their software runs on older devices, even though said devices were perfectly adequate, the latter requiring the use of a new CPU to run the latest version of Windows 10, again for no good reason.

This type of behaviour, along with vendor lock in, constitutes rent-seeking and is harmful, for all the reasons rent-seeking is harmful, both to the economy, consumers, the environment and competition.

And this is one case where regulators should step in. In the EU the warranty for defects is mandated to be at least 2 years. But there is no mention of support periods for dependent products. Regulators require either longer term support for products sold (explicitly setting minimum terms to the pre-sale specifications put forth by the vendor) or opening up of full, detailed specifications for all parts, components or mechanisms involved in the product(s) involved to the public at no cost for the community (or competitors) to maintain.

As long as the operating framework remains unchanged there is little to no incentive for any vendor to change their ways and start supporting their devices for the long haul.



Ζήτημα αξιοπιστίας

Ένα από τα μεγάλα προβλήματα στο ελληνικό επιχειρείν είναι αυτό της αξιοπιστίας. Συχνά, κατηγορούμαστε, ως λαός, πως αποφεύγουμε το ρίσκο, πως είμαστε καχύποπτοι και συντηρητικοί στις ενέργειες, τα πλάνα, την γενικότερη δραστηριότητά μας. Και σε έναν μεγάλο βαθμό είναι αλήθεια.

Υπάρχουν όμως και αρκετοί λόγοι που συμβαίνει αυτό: η ελληνική πραγματικότητα είναι γεμάτη από περιπτώσεις που απογοητεύουν και χαρακτηρίζονται από επιπολαιότητα, πλήρη αδιαφορία για την φήμη ή την εμπιστοσύνη που εμπνέει κανείς, περιπτώσεις ‘αρπαχτής’ από τον πιο μικρό επαγγελματία που θα σου πουλήσει ένα ελαττωματικό προϊόν εν γνώσει του και θα αρνηθεί την ευθύνη, μέχρι την μεγάλη επιχείρηση που θα παραβεί βασικούς κανόνες μιας συμφωνίας ή δεν θα πληρώσει για υπηρεσίες και αγαθά που αγόρασε. Δεν είναι όμως όλοι έτσι και μερικές φορές εμφανίζονται εγχειρήματα που επιδεικνύουν καινοτομία και υπόσχονται πρόοδο και αλλαγή, παρά τις αντίξοες συνθήκες. Και σε αυτό το άρθρο θα ήθελα να ασχοληθώ με ένα από αυτά, μια εταιρία που έχει συγκεντρώσει εντελώς αντιφατικές απόψεις γι’αυτήν και που έχει επιτύχει σημαντικά στην αγορά, αλλά παράλληλα, κατά τη γνώμη μου, έχει κάνει και εξακολουθεί να κάνει πολύ σημαντικά λάθη που ίσως της κοστίσουν περισσότερο απ’ότι νομίζει.



»  Disempowering the user.

I think what really happened was that in the early days of personal computing, decisions were made to give the user an enormous amount of freedom, to communicate without barriers and to share files. And consumers started to use those to, you know, trade information, outside of the boundaries of the law. Since about 2007 or 2008 though we’ve seen a complete shift in this paradigm. Since that time the technologists and the rights holders have really been working together to disempower the user and to turn them more into a customer. So the goal is no longer to empower the computing user, it’s to extract value from them. And I think if you look at your smartphone you’ll see that: it’s a lot more closed than a PC used to be. You almost always have to go through a corporate intermediary. And that was not the case in the early days of computing. There was a period there where the average user had an extraordinary amount of power to do, basically, what they saw fit.

This is a quote by Stephen Witt, author of ‘How Music Got Free’, as mentioned in The Pop Star and the Prophet (around the 20 minute mark), a BBC podcast published back in September — if you’re a music lover in addition to a technology enthusiast, you should listen to the podcast and, perhaps, read the book.

And while his book is probably only tangentially interesting to anyone interested in the history of technology, but without an interest in music, the quote couldn’t possibly be more accurate or well-put.



Δυο ημέρες με το Myo

Πριν από τρία περίπου χρόνια εξετάζαμε το ενδεχόμενο ανάπτυξης του AthensBook ως ένα ‘φυσικό’ αντικείμενο (κιόσκι) το οποίο θα βρισκόταν σε συγκεκριμένα σημεία στην πόλη (π.χ. στο lobby ενός ξενοδοχείου, ενός δημόσιου κτηρίου ή την σάλα ενός καταστήματος) και θα επέτρεπε σε περαστικούς αλλά και τακτικούς χρήστες της εφαρμογής να λάβουν υπερ-τοπικές πληροφορίες ακόμη και όταν δεν είχαν μαζί τους κάποια συσκευή που υποστηρίζουμε (η διείσδυση των smartphones το 2011 ήταν αρκετά — πιο — περιορισμένη απ’ότι σήμερα). Στα πλαίσια της αρχικής έρευνας έκανα μια αίτηση στην Leap Motion για την, ιδιαίτερα καινούργια τότε, συσκευή τους το Leap Motion Controller. Η ιδέα ήταν απλή: να αντικαταστήσουμε το ‘άγγιγμα’ με ‘gestures’ στον αέρα. Να δώσουμε τη δυνατότητα χρήσης της ‘εφαρμογής’ δηλαδή χωρίς την υποχρέωση του αγγίγματος μιας οθόνης (κάτι που είναι, δυνητικά, δύσκολο σε κάποιες περιπτώσεις, όπως π.χ. όταν συναντώνται capacitive touchscreens και γάντια). Λίγο καιρό αργότερα η αίτηση έγινε δεκτή και λάβαμε το Leap Motion SDK και ένα prerelease unit του Leap.

Δυστυχώς η εμπειρία χρήσης ήταν κατώτερη των προσδοκιών. Αφήνοντας τα προβλήματα στη πλευρά του λογισμικού (το οποίο συνεχώς ενημερώνεται και βελτιώνεται) η εμπειρία που είχαμε ήταν αντίστοιχη αυτής που έχει καταγραφεί από αρκετούς, κάποιους περισσότερο και άλλους λιγότερο διάσημους, καθώς και διάφορα μέσα: το Leap είναι ένας πολλά υποσχόμενος, ίσως και επαναστατικός τρόπος χρήσης του υπολογιστή σου αλλά σήμερα δεν είναι ούτε κατ’ελάχιστον έτοιμος για χρήση από το ευρύ κοινό. Τα προβλήματα του είναι αρκετά και πολυδιάστατα: αφενός πρέπει να κρατάς σε σταθερή σχετικά απόσταση το χέρι σου από τον αισθητήρα. Και όσο γυμνασμένο χέρι, μπράτσο και ώμο κι αν έχεις, αυτό γίνεται πολύ γρήγορα κουραστικό, αν όχι αδύνατο. Αφετέρου, δεν έχει καθόλου εύκολο τρόπο να ‘κλείσεις’ τη συσκευή προσωρινά. Τα ‘όρια’ μεταξύ λειτουργίας και παύσης είναι εντελώς αυθαίρετα. Υπάρχουν κάποια gestures που τείνουν να γίνουν πρότυπο αλλά αφενός δεν είναι αποδεκτά απ’όλους ακόμη, αφετέρου πολλές φορές στο μεσοδιάστημα μερικών δεκάτων του δευτερολέπτου που μεσολαβεί ενδέχεται να εκληφθούν κινήσεις και εντολές από τη κίνηση του καρπού, κάτι που σε περιπτώσεις μπορεί να είναι καταστροφικό. Το τελευταίο μέρος σταδιακά βελτιώνεται με καλύτερες βιβλιοθήκες και καλύτερους ‘πελάτες’ αυτών, όμως τέτοια ζητήματα διεπαφής είναι πολύ βασικά και, όπως φαίνεται, δεν έχουν αντιμετωπισθεί συστημικά από το Leap μέχρι στιγμής.

Κατά συνέπεια το Leap είναι μια συσκευή όπου πρέπει διαρκώς να σκέφτεσαι που βρίσκεται το χέρι σου, τι θέση έχει σε τρεις διαστάσεις, κάτι αρκετά κουραστικό πέραν από φυσικά και νοητικά. Τέλος, οι σημερινές υπολογιστικές συσκευές κατά κανόνα δεν ταιριάζουν με το Leap. Ίσως μια μελλοντική συσκευή, απόγονος ενδεχομένως του ‘παραδοσιακού’ υπολογιστή εργασίας (desktop, laptop κλπ), με αισθητήρα μεγαλύτερης διακριτικής ικανότητας, όχι χρονικά (το Leap παίρνει δείγματα στα 290Hz) αλλά χωρικά, που θα απευθύνεται δηλαδή σε άτομα που κάθονται π.χ. πάνω σε ένα ‘έξυπνο’ τραπέζι (βλ. Microsoft Surface — το αυθεντικό, πλέον γνωστό ως PixelSense — όχι η αποτυχημένη-εμπορικά-και-φρανκενστάιν-χρηστικά-ταμπλέτα) να ήταν τελικώς πολύ εύχρηστο. Και φυσικά φαίνεται πως έχει μεγαλύτερο πεδίο εφαρμογής σε ειδικόυς τομείς, όπως π.χ. την μετάφραση νοηματικής γλώσσας σε ‘πραγματικό χρόνο’, κάτι που ήδη γίνεται από προσπάθειες όπως το MotionSavvy, ή την διάγνωση της νόσου του Πάρκινσον κλπ. Λίγους μόνο μήνες μετά την παραλαβή του Leap έμαθα για το Myo της καναδικής Thalmic Labs. Μια συσκευή που επίσης υπόσχεται πολλά, όμως παίρνει έναν εντελώς διαφορετικό δρόμο για να το επιτύχει.

Τι είναι το Myo




The ‘Net As A Utility – Arcs of Political Discourse

In the years between his sensational appearance as a Junior Senator before his colleagues at the Democratic National Convention in 2004 and his election in late 2008, Barack Obama carefully built a public profile of a quasi-radical reformist who, at the same time, is in touch with the world and realistic about the limits and power of politics. He consistently and systematically demonstrated his eagerness to make things better and fairer for the average citizen of his country, the economy and the world. You could argue, and many have since the early parts of his campaign in 2008, that Obama vastly over-promised. His campaign for ‘CHANGE’ had the power to convince even the most disenfranchised voter, the apolitical class.

Just a few short years later, Obama’s pre-election rhetoric seemed as far from his actual policy as his predecessor’s posturing about the US’s military successes in the Middle East. What happened? Being elected to office, I guess; Large-scale domestic and international SIGINT/eavesdropping, Guantanámo, the current mess in Iraq and Afghanistan, his back-pedalling on environmental policy. The US Economy is arguably in a better place than it was when he took over, but he hasn’t managed to revive it to an extent that would safeguard it for the future, while inequality is still a massive social issue in the US.

Over-promising inevitably resulted in under-delivering and this could not possibly exclude issues like regulating the Internet, that are somewhat less important, at first glance, than healthcare or war.

Obama’s policy on the openness of the ‘Net was, effectively, the lack of any policy. The FCC, lobbied intensely by the few monopolies that rule both backbones and the last-mile in the US and the affluent stakeholders in those oligopolies and monopolies, failed to provide any safeguards for the open Internet; in 2013 we saw the first open dispute between a technology company (Netflix) and cable providers when the former saw the performance of its service degrade to the point where it threatened its survival. Netflix ended up paying those providers to make sure its videos reached consumers in reasonable speeds and the end of the open, fair, Internet had begun.



Unwarranted Takedown

A few days ago Microsoft, in what is probably the silliest action they’ve taken in a while now, took down 22 domains belonging to dynamic DNS company noip.com. We know ’cause we use their services at Cosmical. Their move, against a service provider of this sort, is unprecedented and somewhat dangerous from a legal perspective; their argument was that hosts using the dynamic DNS services of noip.com were spreading malware and engaged in illegal activity.

Instead of engaging with noip.com to disable those hosts and block those accounts, they opted to go to (US) Federal Court and get warrant seizing the domains and crippling not just the culprits, but effectively thousands (if not millions) of noip.com customers.

The problem here is that there is no proportion in Microsoft’s response and no concern for the legitimate users of the service, while jeopardising the service provider’s integrity and reputation in the process. It would be akin to disconnecting a nation from the internet, just because there were a few hacking attempts originating in it. By that same logic employed by Microsoft, other service providers, including Microsoft itself, might be in danger of domain seizures, disconnection etc. because a very small percentage of their customers broke the law.

Obviously the responsibility does not only lie with Microsoft here, but also with the Federal Court that allowed and enabled Microsoft to disrupt noip.com’s service. On its part noip.com claimed that Microsoft never contacted them about the problems they experienced and that they would have been able to take targeted measures to stop the abuse from happening without affecting the vast majority of their customers who are now experiencing an outage.

It will be interesting to see how noip.com customers will react; it would be great if a Class Action Lawsuit arose and was filed against Microsoft, which would hopefully lead to other companies choosing more civilized ways of resolving disputes and countering abuse in the future. I guess nothing of the sort is going to happen. Instead of any legal proceedings against Microsoft, noip.com will eventually get hold of its domains back from Microsoft and the service will be restored.

But the very fact that those domains got ‘hijacked’ by Microsoft so easily in the first place and the precedence it sets is frightening and dangerous. That is, the fact that a single corporation, without any due care for the side-effects to lawful customers of a service provider, is able to take the service down, is cause for great concern.


» Fira Sans and Fira Mono

After many years of using Inconsolata Hellenic on my linux and OS X boxes as the monospace font of choice for development, I switched to Fira Mono, commissioned by Mozilla for their Firefox OS and designed by Erik Spiekermann. Inconsolata might have been one of the best looking monospace fonts I’ve ever seen – and the fact that it was free made it an insanely great choice – but it was time for a change. Oh and one more thing, Fira has full support for (monotonic) Greek.


Linode. Affordable, Fast, SSD VPS