Devil's advocate here... but Cyanogenmod building on top of Android is really no different than Jolla building Sailfish on top of Mer, which came from MeeGo basically. It's all about adoption rates - MeeGo didn't get adopted as much, Mer less but they built on top of an existing system. Both of them.

Cyanogen is also, as stated above, quite small too - approximate the same size as Jolla. Cyanogen also is doing something that Jolla is doing - hitting up smaller companies to make their mark (ex. OnePlus, YU, a couple of others) whereas Jolla had Intex (another smaller company).

I think the difference though would be the investors.

The difference here is that Jolla actually funded the Mer project.

Noted. But is that applicable in the beginning during the N8x0 days when Mer was starting out?

The "Mer" from those times has nothing in common (besides the name) with what is known today as Mer.

Well, Mer was mostly product of a person who was Jolla's employee. It was not completely a separate entity and not even remotely as well funded than Android has been for years. Even if Mer inherited some parts from MeeGo (which by the time of deep sixing was not ready), it constitues only portion of what is needed to build an actual device. In comparison, you can take Android, use development tools already there and have complete ecosystem in place via Google which leads easily to wide adoption rates as people really don't have to change anything on adoption. Jolla had to take Mer and Nemo, then create SDK, then do one HW adaptation which went to trash bin when STEriccson pulled the plug from NovaThor, then they coded libhybris, switched to Qt5 and Wayland, do another HW adaptation and put together an infrastucture to support all of it with severely limited funding. So in that regard there is really no comparison between those two.

True. Because Cyanogen is in position where it's easier to succeed. For the reasons I mentioned. Not many are ready to fund Sailfish because if it doesn't fly, they will lose everything and risk for it is way bigger than with Cyanogen/Android. Hopefully they are still able to close the financing round so I don't have to evacuate to Apple-world. At least yet...

It seems silly, but f*ck did they a whole bunch of work!

Note that this was, in fact, part of so-called Zen of Palm

Maybe it's actually on topic to quote parts of it: http://www.cs.uml.edu/~fredm/courses.../zenofpalm.pdf

Let's get to it But first, let me say few basic thoughts:

• it is almost impossible to disrupt the mobile market without huge funding
• this post is going to try to find some ways where a mobile OS could succeed, however impossible it is
• it is always nice to say that you have some differentiation, but you need to have a working system first (like SailfishOS in its current state)
• if Jolla doesn't pick this up, let's hope the community will

I think that currently it is possible to find shortages in current mobile operating systems (let's say Android) in the following areas:

• privacy
• smartness - let the system expect your next step
• using the available resources during idle
• uniform interface
• creative app monetizing

And ideally, connect them all together.

Privacy and smartness

It was around two years ago that I sent (if I remember correctly) an e-mail to Jolla, suggesting a new feature, which would take advantage of the current flow. The feature was "context aware attached pages" - I never got a reply and it was just a pitch, so I didn't expect to get one.

The idea was that when an app doesn't have an attached page, the system could create the attached page according to user's previous behavior. If you were in a menu of an app and last x times clicked on the third option, swiping right would instantly bring you there. I think that this could be currently expanded on - the system could offer this quick action based on the time, weather, ambiance, etc. - all of which could be switchable in Settings.

In order to do so, the system would have to collect some usage data. Currently, mobile devices then ask servers that analyze these data and return the result, which many people do not like, as it breaches their privacy. But what if this data was stored only locally?
Currently, every Jolla does its btrfs balance once a week at night, if it is connected to a charger. But what about those remaining 6 nights? What if the phone analyzed the data collected at night and every day produced new set of expectation what the user is going to need in different situations? This could be done using some basic regressions over available data, with user-selectable variables.

Let go further. I think that one of the best ideas in the Maemo world was Woodchuck - a caching layer that no other other mobile OS has, AFAIK. Together with the statistics, it could make use of plugins (every app could add its own) and every time the phone is on WiFi, it could cache the data user is expected to use. Think Podcasts, RSS, pkcon, Reddit, heavy web pages etc...

Maybe instead of different woodchuck plugins, the system could provide a Qt/QML/C++ element/library that would automatically check for a cached version before getting it from the internet - very much like Kodi does it. It would mean developers would not need to take it into account, but just make use of this system functionality.

Dual monetizing model

On a related note, one big advantage of systems like SailfishOS is how apps adhere to the design guidelines and make use of the default components. This could be leveraged not only with the 'user stats' and 'cache', but also by monetizing apps.

I am also not a fan of ads, but let's think of it this way: what if there was a switch that would add an ad to a Header {} element of Silica and when releasing it through the store, there would be three business models:

1. free apps (no ads)
2. an option between paying or seeing ads
3. paid apps (no ads)

Let's look closely at the 2nd option - you as a developer release an app for $4. If a customer buys it, you get $3 and $1 goes to the Store owner (the developer of the OS). But if the customer decides to go for free version with ads, you get $3 from Store owner and the Store owner then gets all the revenue from ads in that application. At any point of time, the customer can decide to pay his way out of the ads by paying $4, which would then go to the Store owner.

This model would resolve the problem with buying an app without trying it - customers could try it with ads first and then buy it to get rid of them. There would be no model that would force ads onto the user without having the option to pay for the app instead.


Another issue with ads people on Android complain about is how they are data-hungry. But all it takes is to make use of the Woodchuck framework and ads can be downloaded only if a WiFi connection is available (and preferably during night).


That's it. My dream is that one day we will see Woodchuck on SailfishOS, together with a basic prediction what user wants, on a local basis. Let's use the time when the phone is in user's pocket or on a night stand! And if Jolla survives, I hope we will see a monetizing model where ads are an always preventable, but available option.


EDIT: Sorry for being only partly hypothetical. If Jolla never existed, some company would still have to make a working OS first to take advantage of such advanced ideas. So kindly replace the word Jolla with a name of a hypothetical company that would create this OS

EDIT 2: With the prediction of user's actions, I actually implemented something like that in my own app - a helper application for a big film festival - every night, it computes what movies should the user see the next day, based on his preferences and behavior, and it buys the tickets automatically in the morning via SMS. I've been using this app for year and a half, but there was no demand for release, so I just shared it with Jolla users I personally know.

It was biting me for a long time too. What is so special to me with N900?
When I realized it, I summed it up as:

Come to think of it;
was it possible to hook a n900 to a bigger screen and a kbd + mouse?

And also, the whole industry has moved in the opposite direction in the mean time methinks? Would that make it still a device with potential? has anyone thought of getting maemo on the oth yet?