Note to self:
By default, Linux Mint brings OpenDNS (which I hate, if not for anything else, for the NXDOMAIN response it gives) as the resolvconf fallback. Deleting /etc/resolvconf/resolv.conf.d/tail as root and then restarting the resolvconf will (re)solve it. Do not waste time trying to figure out where the DNS is being hijacked on your network: the thing is right on your machine, even if you have overridden the DHCP DNS configuration on the network configuration dialogues – out of the box, it will always use OpenDNS if the DNS servers you or DHCP specified, are not available/don’t answer fast enough.
By the way: http://myresolver.info/ helps with debugging DNS issues.
This post brings to an end months of trouble, and me thinking my ISP, to add to the fact that even though my plan is unlimited, severe speed limiting starts at 15 GB of data usage, was also hijacking DNS queries to their own system. Fortunately, looks like it is not the case. At the same time, I’m disappointed with the choice of DNS fallback by the people behind Linux Mint.
Every news piece I see about Microsoft buying Nokia, seems to focus on the business side of the things – “was this a good move for Microsoft?”, “was this a good deal for Nokia?”; on the future of Windows (Phone), and on the patent portfolio Microsoft just licensed for 10 years (to me, in practical terms this equates to buying…). But I think everyone is forgetting what Nokia did before the Microsoft partnership involving Windows Phone, more precisely the software that runs on their older phones. What will happen to these pieces of software, some abandoned and others not?
I am not a Nokia user and I haven’t had extensive contact with their hardware or software. I do have some friends with modern Nokia feature-phones and non-WP smartphones, and I know a bit about their experience with them. I have also tested some of Nokia’s experimental/in-development Linux based systems on emulators (yes, Nokia wasn’t just Symbian and WP). That said, I apologize for any wrong facts on this post. I will be using Wikipedia as a (improper) “source” for some things.
Nokia developed, or supported the development, of several embedded operating systems. These ran and run mostly on feature-phones, despite some of them supporting third-party software and having other features that make me label them as a “smartphone with alphanumerical keyboard”. S40 is an example of this, having debuted in 1999 and being used until 2012 (it powers the recent Nokia Asha phones). (source)
Recently, Nokia has also developed Symbian-based systems that run on touchscreen smartphones.
The Nokia 7110 (1999) and the 6300 (2007) are both powered by S40
Then there’s Symbian, which was born not just as a Nokia venture. Symbian, a proper smartphone OS, was born as “a partnership between Ericsson, Nokia, Motorola, and Psion” (source). Nokia used Symbian on most of their earlier smartphones, being the responsible for Symbian’s really high market share, which lasted until the end of 2010.
Nokia N80 (2010), running S60
The S60 platform, first released in 2002, was based on Symbian and during the course of its life, it powered phones with and without touch screens (yet again, cases of smartphones with alphanumerical keyboards). Later in 2010, Symbian^3 was released, to power the Nokia N8. More or less by the same time, Symbian, once made open-source, turned into a licensing-model only.
In February 2011, Nokia announced they were dropping Symbian-based systems in favor of Windows Phone. Later that month, the Nokia 808 PureView, with a 41 (forty one) megapixel camera (just like this newer one), was announced, to officially become the last Symbian smartphone.
Nokia Anna and Symbian/Nokia Belle were released later in 2011 as updates to Symbian^3. Some time between Anna and Belle, Nokia outsourced Symbian support and software development to Accenture, a situation that will stay until 2016. The second feature pack of the Belle update is the latest version of Symbian as of now. Will it be the last? I would say so.
Screenshot of Nokia Belle
Nokia also developed apps and internet services for the later Symbian updates (these included an app store, the Ovi Store). Ovi was the brand for many of these apps and services. Among the Ovi apps, there was Ovi Maps, later renamed to Nokia Maps in 2011 and now called Here. Here runs on many mobile operating systems, not limited to those developed by Nokia, and provides map and traffic services to Windows Phone 8. (source)
From Belle on, it’s what you probably already know – more and more Lumia smartphones powered by Windows Phone. For feature-phones like the 2012 Asha models, Nokia seems to have reverted to S40, not Symbian-based.
And this is the story of Nokia’s moderately successful operating systems more or less known by the common user. But Nokia invested in other projects not so widely known to the general public, but which developers know well. It is the case of the Qt framework, which was developed by Nokia after they bought Trolltech, the original developer. Qt was used in Symbian; a bit after Nokia dropped Symbian, they sold the licensing stuff to Digia, that now owns the Qt trademark and the Qt Project.
Nokia also invested in other OS projects besides S40 and Symbian-based systems. Maemo, an open source software platform intended for smartphones and tablets, was based on Debian Linux. It made its first big appearance in 2009, with version 5, as it powered the Nokia N900. From my point of view, this device had some flaws and ended up being only loved by Linux geeks. But this is my personal opinion, of course. I know I would have liked to own one of these 🙂
Nokia N900 (2009) running Maemo 5 / Fremantle
By February 2010, it was announced Maemo would merge with Moblin, to create MeeGo. Finding the story complicated already? The best (or worst) is yet to come… In September 2011, roughly a year and a half after that merge, the Linux Foundation, which hosted MeeGo at the time, announced it would be terminated in favor of yet another project, Tizen. By the same time, Mer appeared as a fork of MeeGo. A company called Jolla, from Finland, then started cooperating with the Mer project to create the Sailfish OS. Confusing enough?
Basically, from the initial two development efforts (Maemo by Nokia and Moblin by Intel) we walked towards six (and counting) different mobile-oriented Linux-based operating systems, some dead, some dying and others just starting. There doesn’t seem to be much app compatibility between these – other than good old Linux binaries (but that applies to Android too and I’m not talking about that kind of apps), and maybe some HTML5 apps (and of course, all the web pages).
In case you didn’t notice, Nokia (now Microsoft) is obviously not involved in any of the current projects (Tizen, Mer or Sailfish) – it’s just not part of their new strategy (and if for some reason anything Linux was involved in the current strategy, I’m sure Microsoft would dump it ASAP, or at least hide it from the public).
Obviously, as of now, one can’t even speak about the market share of these six mobile OS, without having to use scientific notation with negative exponents and a big amount of small-number guessing. To my eyes, Maemo, despite being old and officially unsupported, seems to have a bigger community (and, I guess, a bigger market share) than any of the other small projects that were born from it and forcibly declared as its successors. Or perhaps, I should say that a big part of the community is still centered around Maemo – even though most of the development actually occurs on the new projects.
Add the six I mentioned, to the count of other newborn mobile OS (Firefox OS, Ubuntu for phones…) and make your bets on which one (or whether any of them) will get a market share we can measure, with a percentage above 1%. Maybe if they all got together and united their efforts… oh wait, refer to “MeeGo” for the expected result. But, I’m getting off-topic…
…or perhaps not so off-topic. Nokia, before Microsoft, and during the course of its handset making history, also developed a bunch of different platforms – S40, Symbian and everything that ran on top of it: S60, Symbian^3, Belle and I may be forgetting some. Not all of these were compatible with one another and they had different user experiences. Now compare this to Apple’s iOS since the first iPhone, or Android since the first phones that ran it. Sure, there were lots of updates to each of these systems, and sometimes the user experience changed a lot, but they never changed names and there was always some kind of data and app compatibility.
Applications one developed for Android 1.6 most likely still work fine in 4.3, even if sporting an outdated user experience. The user experience has changed but progressively, and much of the practice an user gets from using earlier versions is still valid on the latest. On the Nokia side, from S60 onwards, there was some app compatibility (but from what I see, not as good as Android’s or iOS’s), but the constantly changing names and looks confuse consumers.
MeeGo also ran on netbooks, in-vehicle infotainment devices and other embedded systems.
Since the Linux-based Maemo and successors had nothing to do with Symbian, Symbian-based and Linux-based systems always lived in different universes, despite being both made by Nokia for similar devices. And now Windows Phone has nothing to do with either Symbian or Maemo/MeeGo – it’s yet another “ecosystem”. Talk about fragmentation in Android, and I’ll show you all the systems created by Nokia.
Nokia managed to start more mobile OS projects than companies like Microsoft, Apple or Google, and yet didn’t manage to stick with any of them, except maybe Symbian… but even on the Symbian land, there were lots of changes to user experience over the time, and to the end user it definitely didn’t feel like it was the same OS when they bought a new phone.
And now that Nokia is basically Microsoft, users of the multiple platforms that reached the mainstream consumer before Nokia made Windows Phones, are pretty much screwed. Nokia/Microsoft no longer cares about their non-WP phones except for what I call “traditional support” (phone breaks, gets replaced or repaired; security bug is found, gets fixed) – but users expect other types of support for smartphone OS, namely updates (to fix non-critical bugs, for example). And maybe even the traditional support is compromised.
I see a lot of wasted work and money on the Nokia side, investing on things that eventually never went forward because at some point they lost support from Nokia. I even think much of the recent waste over the past two/three years was purposeful (*cough* Stephen Elop *cough*). At the same time, I can see the old Nokia strategy as “not putting all the eggs in the same basket” – they weren’t just committed to Symbian, but also S40 and Maemo/MeeGo, and later Windows Phone. Then they decided to drop all the eggs but S40 and Windows Phone. And now that Microsoft bought Nokia, I believe they are going to drop the S40 egg too.
Ultimately, I believe this isn’t good for the consumer. It loses yet some more mobile OS choice, even if these choices weren’t so popular – at least, they were something different. The choice is more and more between Android, iOS and Windows Phone, and if the currently in-development projects don’t hurry to get on the market to earn some users, they may suffer the same fate as all the other ones that got killed or merged, eventually turning them into abandonware.
It would be good if one could go to a store, buy any smartphone or a tablet, and later choose what OS to run on it. Like what happens with desktop and laptop computers, you buy any one of any brand (even Apple!) and you can almost certainly run a OS other than the default on it – some things are taken for granted on a PC, which allows for this kind of freedom of choice.
Sure, you can install custom ROMs on smartphones… now go ahead and buy any one that isn’t popular and see what happens. If there is not a big enough user base most likely there isn’t anyone to develop custom ROMs. And unlike what happens with PCs, there doesn’t seem to be some list of assured features and standards. On personal computers, even on the ones built from OEM parts and no recognizable brand, it’s as simple as setting a boot device that is an OS installation disk. Sure, you need drivers to get the full potential of the hardware. Still, some things Just Work(tm), like PS/2 keyboard and mouse, serial ports, VGA video out (and I don’t mean graphics acceleration) and disk I/O.
Isn’t it about time some sort of standardization starts happening for mobile devices too, both at the hardware (charging connector/method) and software (system booting, basic elements)? The idea would be that simple things like storage I/O, screen (without GPU), touchscreen input (even in single-touch mode) and hardware keys Just Work(tm) because they are really easy to drive by the OS?
Look at something like KolibriOS. It is pretty limited on what it can do on a PC, because of a limited amount of drivers are included, and anyway everything is very simple. But as things are now, there’s no way one could write a KolibriOS that would run on all the phones and tablets the same way Kolibri runs on almost every PC, even with limited functionality.
It’s certainly a distant dream, seeing that even charging connectors are something we are yet to get quite right… all to make big brands happier at the cost of unhappier customers – even if they never noticed they were less happy because of that. </rant>
If I tried to crowdfund the renewal of my domain names or server hosting, I wonder how well that would work…
For a long time, I’ve used YOURLS in my URL shortener projects. I have always liked extending it, so that it did something more than just URL shortening. The results of my work have turned, over the past two years, into what tny.im is today.
Until a few days ago, tny.im was running with software based on YOURLS. Yes, “based on YOURLS”. It wasn’t running “on top of” YOURLS. As I added more features to this URL shortener, I found it easier to just modify the core files and add rows to database tables at will. This came with a price: updating YOURLS without losing my modifications and while keeping database compatibility was really hard, requiring me to rewrite all the modifications. To make things even worse, I had modified some of the core files for them to work with the Bootstrap CSS and JS things. The statistics page (yourls-infos.php), which I had also managed to modify to a point where not only it was Bootstrap-themed, but was also the main UI for users to edit short URLs, was a specially hard problem to solve if I was updating YOURLS.
I could have stayed running the 1.5 version of the said script forever. However, it lacked many under-the-hood improvements of the 1.6 versions, and as new versions would be released, it would only become more obsolete. Again, a big problem resided on the link statistics page: it wouldn’t handle links with many clicks properly, because it would use a lot of memory. This bug no longer existed on 1.6, and this was something that kept me thinking I really should update to a newer version.
Another thing that I had heavily modified for tny.im was the public API. I had modified it to not support some methods which would disclose too much information (e.g. long URLs for paid-access links and links which had reached their hit limit), as well as added support for tny.im-specific features like the passcode, hitlimit and Bitcoin related things.
My code for all these things was really ugly. It had been progressively added and changed over the course of two years, had multiple coding styles, multiple indentation styles, multiple bug styles and God-knows-what-else. The fact that I only worked on the project sporadically meant that I often didn’t remember what I had done already, meaning there were giant mistakes like two variables for the same thing. Somehow it all worked, well enough to take to the correct destination over two hundred thousand short URL clicks.
Some days ago, I finally decided it was time to do something. I wanted to add new features to tny.im and future-proof it at the same time, but the code was impossible to maintain – the original core code, mixed with my bad code, made it seem it would inevitably break if I touched it. I knew, from the start, that the proper way to add features to YOURLS was to code plugins – but laziness, convenience and the fact that version 1.5 didn’t offer that many plugin hooks made me modify the core files, as I said above.
I had to start from scratch, doing things in a “staging” vhost that had no communication to the live tny.im website. I started by installing YOURLS on that vhost, on a database separate from the live one, of course. Then I accepted the challenge of trying to implement all the tny.im features, including the Bootstrap theme, without ever touching the core files of YOURLS.
Fortunately, version 1.6 of YOURLS had many more plugin hooks. But I knew I wouldn’t be able to implement every feature as a plugin. Things ended up like this: I consolidated my hackish plugin soup from the old tny.im scripts in a single “tnyim-framework” plugin, and things like the index page UI, link lists/folders, Bitcoin address shortening and “internet toll” features, as well as login/logout (which is separate, and has always been, from YOURLS auth methods) would keep being separate from YOURLS.
Instead of modifying core files like functions.php and functions-html.php to add my code, I put them in a separate tny.im-specific folder. It has it’s own “load-tnyim.php” file, in the style of the “load-yourls.php” file, which loads the necessary variables and files for the tny.im features and UI.
The problem with the statistics page, as well as some modifications I had made on yourls-loader.php, was solved my creating tny.im-specific files that are similar to the YOURLS ones and perform the same functions, but have my modifications. The new statistics page is based on the YOURLS 1.6 one, and it was a hassle to modify to meet the functionality of the old one, that is, Bootstrap-themed, and with that “Manage” tab that allows people to edit links. This was mainly because, as @ozh says, it is an “awful HTML/PHP soup”. Also, it has little to none plugin hooks from what I can see, but even if it had, they would never be enough to allow me to customize it to the point where I did.
Things like link preview, and the API modifications, went in the tnyim-framework plugin. And the reason why I think YOURLS is awesome, is that I managed to change much of the API behavior, adding new return and request fields, as well as obfuscating some, just with plugin hooks.
As for the database, rows specific to tny.im are now kept on their own table. The functions.php on the tny.im folder has methods for adding and editing URLs that handle the tny.im features. It works like this: the code calls my custom methods for adding/editing short links; then, things are added/edited on the normal YOURLS DB table using the core methods, and then my code takes care of the other table which stores things like link hit limit and price (for internet toll links), as well as adding the link to the users’ account when appropriate.
Speaking of user accounts, the data that relates users to the keywords they have access to, is kept also separate from YOURLS tables. Same with the lists of links feature.
Finally I moved the columns of each database table to their correct destination in the new tables – five hours worth of SQL commands. Then I moved the new script files to the vhost of the live site, after editing the YOURLS configuration file and the nginx vhost config of course. It seemed to start working right away. All this took me two days.
Now the code for tny.im is much more clean, readable and most importantly, maintainable. I can finally add new features to tny.im without breaking half of the existing ones. And I can update YOURLS without breaking the whole thing, since none of the new features are implemented in the core files. @ozh’s improvements in pages like the links statistics one and yourls-loader will not automatically get merged with the tny.im code, but I can add them manually while keeping my changes.
I’m not yet fully sure the new tny.im is clear of bugs, but over 90% of it seems to work the same or better than before. It should look and work pretty much the same as it did before – which only proves, that I really made a wrong move when I started editing core files, since everything could have been done without touching them. I think I learned the lesson, hopefully not only for YOURLS, but for most scripts which have plugin interfaces.
I think I made the most full-featured URL shortener ever seen, and obviously I’m proud (certainly too proud) of it. It’s all built around YOURLS, and I bet you wouldn’t even tell it was powered by that script at a first glance. Stay tuned as more features and bug fixes are to come.
Six months after the previous public release of my Utilities add-in for the Casio Prizm calculators, the new Beta 9 is out! This version comes after an important refactoring of the source code, which by itself led to many fixed bugs and much higher stability. The g3a binary is also smaller (about 136 KiB) while featuring the same features plus some new ones, and that’s a big plus on a device with limited storage memory like the Prizm.
The most noticeable new “feature”, is probably the increased stability on tasks and calendar events, as well as in the file browser.
You no longer need to keep the storage memory fully optimized for calendar events and tasks to work without system errors.
Another new thing is a new timer/stopwatches system, which is now called “Chronometer”. There are 20 chronometers and each of them can go upwards or downwards. You can now select multiple chronometers (hint: Shift+F1 on the chronometers list shows useful selection options).
There’s now also a way to open files as text, with automatic line ending detection (supporting CR+LF, CR and LF) – on the file manager, press EXE when hovering on a file, then press F1 on the file information screen to open the text viewer. This still doesn’t automatically convert special characters from ASCII to Casio’s multibyte system but it’s something I’m planning to add.
File copying has been added, but this is unfortunately still too unstable, so this option will only appear when the “Show advanced tools” setting is enabled. Note that this is one of the last two functions of the add-in which are unstable (the other is the Add-In manager, which keeps being unstable probably because of a OS limitation – one can’t touch the add-in array under normal circumstances…).
Now there’s a way to save the current time to a OS Basic variable (those you can use on Run-Mat and almost everywhere in the official software): press the X/θ/T key on the main screen (where you see the time), and the hour fraction will be saved to the variable T with the maximum precision the RTC allows.
This add-in is finally headed for a stable 1.0 release – this Beta 9 release would be called a RC, if it weren’t for the Add-In manager and the file copying function, as well as a little feature that was committed to the git repo already in March, but got lost as I refactored the code. I shall add that little feature in time for the next release.
(only applies if you used Utilities before Beta 9)
This bit is really, really important as there are some folder names that changed with this new release; also, not all files are compatible (calendar events and tasks are compatible, though). You must follow the following instructions for Utilities Beta 9 to work properly:
– Delete the @UTILS folder in Main Memory (Memory app in Main Menu, press F1, scroll down to @UTILS, press F1, F6 and then F1);
– Now connect the calculator to your computer through USB, as a flash drive, and perform the following steps:
– Rename the @CALNDAR folder in the storage memory of your calculator to @UTILS (you must do this on the computer because the OS won’t accept the @ as a valid character);
– Inside the newly renamed @UTILS storage memory folder, delete the file Hash.plp in case it exists (the new Hash.plp file is no longer compatible with the ones from previous versions).
– Copy the new utilties.g3a (download below) to the root folder of the calculator, overwriting the old one.
– Safely disconnect the calculator from the computer. You are ready to use the new version once your calculator finishes updating the Main Memory…
Download Latest Version of the Utilties add-in for the Casio Prizm
That moment when Casio decided to call their fx-CG 10/20 series Prizm (with a Z) instead of Prism, they couldn’t imagine how much of a bad PR they had managed to avoid…
(I feel sorry for all pyramids and alikes out there…)
If anyone knows of a instant messaging and VoIP solution that’s free, privacy friendly (end-to-end encryption, preferably) and, most importantly, easy to get the “average Joe” to switch to (i.e. my non-technical friends), please let me know. Bonus points if it has interoperability to work with other systems (at the eventual cost of losing some of the privacy) – that way, I wouldn’t need to get all my friends to switch.
(Of course, Google Talks/Hangouts, Skype and anything Facebook-related is out of the equation.)
Another thing I’m looking forward to is having my computers (Linux and Windows), old Windows Mobile phones and my Android phone and tablet synchronized (contacts, calendar, email) through the Internet without Google’s help, but I realize this is even harder find than the IM solution I specified above.
Seeing that I’m getting a lot of hits on this blog from people searching about the SliTaz servers being down right now, probably because I have a blog post with the words “SliTaz” and “dead” on the title, which was published over a year ago, I’d like to inform that I know nothing about the server situation other than the fact that the website is unavailable for me too.
(talking about SliTaz, I kinda stopped following the project, even though I find it interesting… note to self: pay more attention to it from now on)
I moved this blog to a different server (no downtime because the transfer was planned). Please report if something is broken.
On other topic, I have been very busy with school and real life in general, even on weekends; that sums up the reasons for my recent online absence.