Electric cars for the rich?

Norway is a small country off somewhere above the top edge of the map. But there is one respect where it is huge. That is in the context of electric cars. In 2019, there were more than twice times as many "plug-in" cars (electric vehicles, EVs) than petrol/diesel cars. This is of course the result of government policy over many years, to promote "zero-emission" cars. The target is that in five years, 2025, all new cars sold in the country should be electric.

Controversy: are EVs bad for you?
But of course, that costs money. The loss in revenue for all EV measures has been calculated to some €200 million a year, and many see this as a subsidy from tax payers to EV owners. It has raised some controversy about whether this is justifiable, both economically and environmentally. The Michael Moore argument ("Planet of the Humans") that we should not support EVs because they are blacker than petrol since electricity comes from coal, does of course not wash here - we are 100 per cent hydro-electric - and is not true in the US either, I believe. But it highlights what is an actual dilemma: EVs are of course not fully green, because they have to be made, and it is true that more CO2 is emitted from producing an EV than a diesel/petrol car ("internal combustion", ICE)  primarily because of the battery. And longer range means automatically larger battery (not to mention the 2,5 ton e-Tron with its 95 kw battery, more than twice that of a new Leaf). So, there is a dilemma here. The very popular long-range EVs help the transition from petrol to electric. But at the same time, they are less environmentally friendly than the more modest cars which you may have to recharge more often because they have a smaller, but lighter battery.

EV batteries - they are bad for you

There was an argument in a web comment recently, which I thought was interesting to look into. It said, in essence, "... and you cannot defend, in environmental terms, to replace a petrol car that has 10 years life ahead of it, with a new EV" (because of the production emissions). That argument sounds plausible, but would it hold up? Now, a car with 10 years ahead of it (meaning, statistically about 8 years old) may well find a new owner, but let us ignore that and make the calculation simple: to be justifiable environmentally, the cost (in emissions) of producing the new EV must be lower than the CO2 that it replaces, i.e., that would be spent by driving the ICE for ten years. This production cost is in fact surprisingly large: just under 10 tons CO2 (for a car weighing a bit over one ton - the example is based on a 2019 Nissan Leaf). But an average ICE driving the average of 13,000 km a year, emits about 2.75 tons CO2 annualy. So the EV has saved its production cost already after about three and a half years - the argument is false. Nevertheless, for green purposes, the battery issue is definitely something that should be improved, by changing how they are produced, or improving reusage and recycling of the rare materials in the batteries. Of course, that is still a few years ahead for the current surge of EVs.

The social dimension: Should we subsidize electric cars for the rich?
The above argument is general. But we have also had a discussion from parts of the left wing in this country, which are irritated that wealthy Tesla owners get their toys subsidised by tens of thousands of euros of tax-payer money. That the green left supports this, shows their class affiliation, they whisper: the latte-drinking EV owners who have never set foot on a factory floor. The mentioned latte drinkers just sneer back. Yet, the argument is worth considering. It is based on these claims:

(1) Whatever you say, normal people do not drive EVs. They drive petrol or diesel cars. This is quite true. In spite of the surge in EV purchases these last two-three years, about 90 per cent of kms driven are done by petrol or diesel cars, statistics show. Also, the critics follow up: People with average income do not buy new cars, they buy used cars. Only the richest buy new cars, and used car sales are almost completely for ICE cars. That is also demonstrably true, about three of four cars sold are second hand, and they are overwhelmingsly petrol or diesel. But that is of course just an effect of the number of used cars there is on the market. The surge in EV adoption only took place a couple of years ago, and they have yet to replaced by their first-time owners. If the EV share of new purchases remains stable, used cars will eventually become EV or ICE in the same proportion as new cars today.

The red line is the market share.
Source: Elbil.no, from Public Car Registration

      However, that raises the reflection of how EV used car prices will hold up. All EV supporters claim that EVs have less wear and tear and require less service than ICEs, and will therefore keep their value longer (good for sellers, bad for buyers). On the other hand, it must remembered that an EV is basically a computer on four wheels. And we upgrade our PC not because it breaks down, but because the new one gives us functions we need or want. A new petrol car is more efficient, safer and better than a seven year old one, but it operates basically in the same way: It is mature technology, and functionally, you will find the old one just as useful as a new one. EVs are still in rapid development; every year introduces drastic new changes in functionality (as in battery management, e.g.) which may make a seven old machine dated almost like a seven-year old PC running Windows 7. Granted, an EV is a bit more expensive than a PC, so you would want to hold on to it longer, but the point is that this may lead to a more rapid decrease in value compared to a new one. Which is good for the buyer, and will help removing any "rich man's toy" image of normal, regular EVs.
      But anyway, in order to get EVs into the used car market, they first have to enter as new cars. So, the social differentiation we see today will disappear gradually over the coming few years.

(2) Nevertheless, we do know that even among new cars, the share of EVs follow the income distribution. A statistic (from 2018) shows that among new car buyers, the 25 per cent with highest income ("the wealthy"), 56 percent bought EVs, while among those with the lowest income, only 33 percent bought EVs. Furthermore, ICEs are more used than EVs, they have a higher number kms driven. Now that is very interesting, when we just postulated that a new EV has a similar or lower cost than a comparable petrol car. So economically, it should make more sense for the price conscious to go EV. Why do fewer in that group do so?

A car for the rich, paid by our tax dollars?

      The critics say this shows that the wealthy can afford to have two cars. They use an ICE by preference, but can afford to get an EV for the commute, in order to drive in the bus lane and escape the road tolls. EVs are just a toy and a status symbol for the rich. Teslas, gaah.
    It is also true that left-leaning people may gag over the number of Teslas on the streets of Oslo, because there are indeed a lot of them, and they are of course quite noticeable. Of course there is a status symbol show-off element here. However, it is not Teslas that actually dominate the purchase statistics, but normally-priced EVs that just look like other cars. You dont't notice them on the street unless you look closely.

Or just another car?

     It is more likely that these figures show a quite  different reality: People do not buy a second car because they are rich, but because they aren't. Two-car households are generally two-income households, i.e. families, and they more often than not have to calculate before they get the second car. They certainly could not afford to spend €20-30.000 on an EV just to save €2-3 on a road toll, that just does not make economic sense - particularly as these perks are progressively reduced. But families mostly do need at least one long-range car for holidays etc., and EVs (non-Teslas) have until these last couple of years not been seen as sufficiently practical for that. So of course their ICEs will get more mileage: Both cars are used equally for daily commutes, one by each spouse, and only the ICE will be used for longer family trips in weekends and holidays.

This may also help explain the sudden surge since 2018, in spite of the prospect of reduced perks: What is new in these last couple of years is the increased availability of models that may be called "affordable long-range" EVs. We always had the Teslas whith their 400 km or 500 km range, but at prices out of reach for most people. Now, there are several models with a range of 400 km or higher at a price below €35.000, going as low as the Renault Zoe (390 km) currently at €23.000. And, if we look at the best-selling models, these are precisely those that top the statistic: Zoe, Kona, Leaf, Ioniq (Soul and e-Niro were unavailable, but are in the same category). So, what we may be seeing is that more single-car households are now accepting such models as a realistic alternative for a car that they can also use for the longer trips, even though there are still issues left that favour combustion engines. If these practical aspects are solved in the time to come, these single-car households with moderate income will also increasingly choose an EV over an ICE. 

     Plug-in hybrids (PHEVs) are also an alternative as "both commute and weekend" car, but the market share for PHEVs has stood still around 15 per cent since 2016, and actually took a small plunge in 2019, supporting the argument that long-range full-electric has taken over some of that segment in the last two years.
      In other words, what people want and will favour, are not flashy or star-in-the-eyes EVs, but practical tools that "can be used in the same way as a petrol car", just a normal car, at a price comparable to an ICE car. If we, as the critics suggest, cut the subsidies to raise EV prices by 50 per cent, these families would not drop the second car, they would buy two ICEs. Thus, the EV subsidy policies do favour ICE to EV replacement, which was the intention, and are to the benefit of normal families.

That does not deny that Tesla owners absolutely do benefit from the subsidies. But then the intention of the EV subsidies is not social equalization, it is to move from petrol to electric. A rich person will remain rich, and if we price a Tesla out of his range, he will not buy an e-Golf. He will buy a prestige ICE, a Porsche or similar. He may have bought a Tesla just because it is cool, that is all right, he will still help with the green change, whether that is his intention or not. The point here is that it is less and less true that EVs are cars for the rich. When they have reached more than fifty per cent adoption, they will be just cars like other cars, for normal people who drive cars.
       That said, it must of course be remembered we are talking about cars against cars. Clearly, both for green and red purposes, it is better that people do not drive a car, but use public transport, or walk or bike instead. All cars, including EVs, congest the roads, wear down the road surface to whirl up particles and require parking space. So EV subsidies should of course always go in tandem with increased financing of public transport and bike lanes. But those are seldom in competition, the political forces that support green cars generally also support green transport generally. More funding to both, and less to polluting!

The un-equalizer - not the car, but the garage?

However, there is another factor which is not so often considered and has a social aspect. In order to have a plug-in vehicle, you must have somewhere to plug it in. The most practical, and economical, solution is to plug it into a home charger which you have in your garage, or on a permanent parking space where you can install it. But many people who live in apartments do not have a garage. That does constitute a social differentiation. So, to combine zero emission conversion (for those who need cars) with social equality, this is an issue that should be addressed, so that everyone has equal access to EV as to petrol cars including low-cost charging, irrespective to what kind of house or flat they live in. That might become even more important as EVs begin enter the lower-priced used car market in a few years, so that you can choose practical EV cars even if you are in the lower price range. 

Incidentally, at the end: This last bit also concerns the idea that EVs are "urban vehicles", which the old 100-km EVs mainly were. But in the countryside, most people do own the property where they keep their car and can set up home charging. They would hardly need access to public charging in their own village, at least with a long-range car. So, in fact EVs may be easier to adopt outside the city than inside. But again, the type of car may be important. Pickups aren't that much used in Norway, but trailers are, at least for someone working on a farm, and none of the regular low-to-medium priced EVs today allow a trailer to be attached. That may be a marginal issue in the city, but possibly a priority for people who want to use their car for their farm work. So, we can only hope the producers begin taking that into account and focus on making practical family cars for the same kind of use as any other car.

In short, walking is best, biking is excellent, buses and trains preferable, but EVs are also a good thing for the reds as well as the green in the political landscape. 

Travelling green(-ish) at home and abroad

This blog has been mostly moribund over the last few years. It was established to talk about Macintosh matters, but I have let it slide. Maybe because there weren't that many entertaining crashes and collapses (oh, yes there were - like when I thought I could run my iMac from the hard disk only, and had to wait an hour just for it to boot. Or the mysterious Catalina bug that reportedly could delete gigabytes of mail). Anyway, it didn't happen and as you can see from the posts, I have on occasion also written about other stuff, like radios and ugly news sites.

But, since the blog exists, perhaps I could redirect it more widely to matters technological in general that happen to interest me. Maybe the reader is not interested, but I will just write what falls into my head, that is how the blog works. Mac stuff will still be there, but alongside whatever else I find interesting. But my approach will still be that of the fairly regular user, not the nerdy stuff I do not understand anyway. I am a normal professor of history from Bergen in western Norway, and this will be the world I see from my window.

Travelling green without going overboard
So, for the first instalment this year, I thought I would share some musings I have about choices we, both academics and others, will have to start making. As I write this, the world is locked down, borders are still closed and the air is cleaner that it has been for decades. But that, alas, will pass. Well, alas for the last part, the clean air. In a few years will look back at this time with the same nostalgia as the 1973 oil crisis, when the streets were also empty because driving cars were restricted or forbidden. We will be travelling again some day.

But the need to cut down unnecessary travel is something that has been discussed more seriously over the last years, and we have been asked by our institutions to reduce in particular air travel, for environmental reasons. I decided already a year or two back to refuse to go to Oslo just for a single meeting or one-hour examination, and demand videochats instead. Now, that is becoming even more practical since we all have had occasion to practice on Zoom and similar.

West Norwegian airplane

But we cannot stop travelling altogether, of course. It is just that when we do, we must, in addition to airfare prices and time wasted, also consider means of travel and environmental impact. When can we use a train, and when can we justify taking to the air? There are choices to be made, and your "green-ness" may direct your answer. I am now reflecting about my own choices, as a "reasonably green" person who accept a certain level of pragmatics, to go fully Greta Thunberg with a sailboat to America is beyond me. Your choices may differ from mine, however, I thought it might be useful to get some figures on the table to see what exactly we are talking about. What are actually the emission cost of travelling by this way or that way? Is it better to drive a car than taking a plane? Do propeller (turboprop) planes as we use locally in western Norway spew out less CO2 than jet planes (yes, they do, and by a pretty hefty margin on short hops). And, how much extra time will it actually take me to travel green?

So in a free moment, I started lookng at figures for my comparison. There are quite a few sites around that count carbon costs, and although they do not all match, and are of course based on averages and presumptions - clearly, the more people you put into a plane, the less is the emission per passenger (and did you know that narrow single-aisle planes are better than two-aisle planes?). However, I was able to set up this list, which is more or less consensual among the calculators:

Grammes of CO2 emitted per person per km travelled:

• Jetplane, international: 195 g.
• Jetplan, domestic 270 g. [much higher, because more fuel is spent on takeoff and landing, than on cruising]
• Propeller plane: 95 g.
• Car (average, diesel engine [ICE], driver only): 171 g.
• Bus / long-haul coach: 27 g.
• Train: 14 g. (West European average. In Norway, almost all trains are hydro-electric, so 0 in this country)
• Ferry w/car: 128 g. [passenger only: 18 g.]

Travel strategies, from Bergen to the west country
So, I began setting up scenarios and I how I would answer them. The presumption is a business / work trip for one full day away, and one person only (not holidays). For such a actual work period, I would imagine that spending 2-4 hours extra to travel green is quite reasonable, while spending a full travel day (8-9 hours) extra each way for one day at work would be excessive. Here, I do not consider economics, just time vs. CO2 emsissions. How do the destinations stack up?
      And here I must apologize for being very Norwegian. I write this in English, because the blog is in English, but probably those who know a bit about Norwegian geography will be at an advantage here. But hopefully, the basic thinking beneath this might be transferable to other situations.
   
Destinaton one: Bergen to Oslo.
Distance by car: 460 km. Options: 

• Plane: 81 kg. CO2; time spent 4 hours town centre to town centre [For planes I normally add 2 hours ahead of departure, and 1 hour after arrival]. 
• Car [ICE]: 79 kg. / 7 hours [Google maps suggestion] 
• Train 0 g / ca. 7 hours. 

      Sum: As Oslo can be reached by train, as the only destination discussed here, train will always win. The extra time expended is 3 hours, which is fully acceptable. There are even night trains, but depending on schedule, you may have to add a hotel night. But travel to Oslo will be by train.

Destination two: Bergen to Volda
[Volda, north of Bergen, has a college we often visit]
Distance by car: 352 km. Options:

• Coach bus: 10 kg / 8 hours
• Plane: 22 kg / 4 hours [Propeller plane]
• Car: 60 kg / 6-7 hours (depending on ferry)

      Now this is a rather more interesting example. The ICE car is clearly out, it emits almost three times as much CO2 as the plane! When I went last year, I found no direct bus, only an indirect one with a night stopover. So, then the plane was actually the best alternative. Now, the coach adds 4 hours - the limit I set for reasonable time waste - which makes it basically a full day's travel each way, and emits about half the CO2. Is that worth the extra time?
      However, we live in Norway! As everyone with any green awareness will know, Norway is a different planet, where everyone drives an EV (electric car)! Well, not quite. But electric is indeed an option for those who own one, and a clean one (0 emissions on our hydro-electric grid). The car can take a bit more direct route, so the time cost can be 2-3 hours only, and in particular gives you greater flexibility than the once-a-day bus. Still, you must factor in at least one night's extra stay over the plane, probably two with the bus. It becomes more of a toss-up between bus, plane or whether you feel up to driving for 13 hours over two days.
   
Destination three: Bergen to Førde 
[a regional centre, on the way to Volda]
Distance by car: 175 km. Options:

• Bus: 5 kg / 3,5 hours
• Plane: 11 kg / ca. 4 hours
• Car: 30 kg / ca. 3,5 hours (depending on ferry)

 Førde, being closer, has a much better coach connection, like the plane three times a day, and as we can see, actually beats the plane on time spent, so here there is no contest, bus wins. But for an EV it is also well within reasonable range, and with less emissions. So either EV, if you have one, or bus.
   
Destination four: Bergen to Stavanger, 
our neighbouring university town to the south.
Distance by car: 210 km. Options:

• Bus: 6 kg / 5,5 hours
• Plane: 42+ kg / 4 hours
• Car: 35 kg / ca. 5 hours (with ferry)

 Bergen-Stavanger is one of the busiest air routes in Norway, and the one it is most urgent to get rid of. The CO2 sum I calculate here is probably far too low, since the plane basically just takes off and then lands. They are even planning some humongous bridges and tunnels to make the road alternative competitive, but as we can see it already almost is, in spite of its 45-minutes ferry stretch. Buses are frequent, and provide the best alternative besides an EV. So here too, bus or EV. Here even the ICE car marginally beats the plane on emissions, but definitely, take the bus.
   
So, the conclusions for western Norway are not terribly surprising, except for Volda, where the argument for flying is actually quite realistic.

Now for a trickier one. The third largest city in Norway is Trondheim, in central Norway.
Distance by car: 629 km. Options:

• Bus: 17 kg / 15 hours [overnight bus?]
• Plane: 115 kg / ca. 4,5 hours
• Car: 107 kg / ca. 10 hours (at least one ferry)
• Train: 0 kg / best case about 15 hours. 

It is actually possible to take the train from Bergen to Trondheim, by a detour through Oslo. The bus and car cut straight north-east across the mountain, by different routes evidently, the car is five hours quicker, but the ICE car emits almost as much CO2 as the jet plane. Anyway, the point is that for all practical purposes all three options give unacceptably long travel for a one-day meeting, it would take two full days (or a night and a day) each way, unless you are young and sporty enough to stomach an overnight bus journey both ways. Even a night-and-day train would only get you there in the afternoon of the following day. It is way above the 4-hour limit I set. So here you will see the fortitude of green conscience: A strong supporter will take the train and spend the extra days, a lesser one - and I fear I would be one - would, if I cannot get them to do a video meet, take the CO2 cost and fly.
      Further north, even the staunchest green supporter will acknowledge that flying is inevitable. You can, by spending about yet another day, get a bit further north by train, to Bodø, but the rest is bus and ferry, and perhaps only local buses. The most idealistic young people will take the train from Bodø, or from Narvik (through Sweden), but a car, even an electric one, would only be for a holiday trip, not for a business meeting.
   
Travelling to Europe
Again, so far not so many surprises here: use a bus if you can, reduce flying to a minimum, but fly if you must.

However, the most pressing issue is travelling abroad. We can and should not completely stop going abroad, and it has been my contention that it is a practical impossibility to go from Bergen to the European continent without flying. So let us test that claim with an imagined visit: Let us choose Paris, which is pretty far north in Europe, and assume we are going for a guest lecture - a morning's work with a lunch or similar. How long does it take us to go from Bergen to Paris for that purpose without leaving the surface?

Option 1: Train
As far as I can see, there are three main options. One is by train from Bergen to Oslo, then down to Copenhagen, and from there through Germany to Paris. I did that many times in my youth (from Oslo, but still). But that was then. Now, the fragmentation of train services seems determined to deter us from travelling long distances. There are hardly any night trains left, and the train you must take leaves an hour before the one you are on arrives. So, trying to use various scheduling services, I came up with this travel plan:

• Day 1: Leave Bergen on the night train, arrive Oslo in morning, after the early train to Gothenburg has left.
• Day 2: Take the noon train to Gothenburg, change and continue to Copenhagen. No night trains going south, so hotel night.
• Day 3: from Copenhagen to Duisburg in Germany, arrival 8 PM. No night trains.
• Day 4: Duisburg to Paris, arrive noon. [Once a week, an early train will get you all the way from Cph to Paris in one day]

Total time spent: 33 hours effective travel, or 2,5 days and three nights. CO2 (continent only), about 17 kg.  
       (There is another option: leaving Bergen at 8AM on day 1 will get you as far as Helsingborg at midnight; then wait there until 5AM and continue to Malmö and Copenhagen. That will save you a day, but is possibly not an attractive option).

Option 2: Driving through Sweden (in an electric car)
I will not even calculate the CO2 of an ICE (well, I did, it was 390 kg - a plane emits 260 kg). But you can drive EVs in Europe, of course, if you have one. There are two main options. One is to follow the train route, to Oslo, down Sweden and Copenhagen through Germany. That will take you effectively 26 hours travel, and depends how many hours you like to drive a day. If we set a limit of 10 hours / day, that will then include two overnight stays on the way. CO2 is not 0 now, as we must calculate in non-clean European electricity for that part of the journey, estimated at about 40-45 kg.

Option 3: Driving through Denmark
The third option is to take the ferry directly from Bergen to the northern tip of Jutland, and then drive down Denmark and Germany. The ferry costs 72 kg CO2, leaves in the afternoon and reaches Denmark the following morning. You then have fourteen hours of travel to Paris. So, a total of 35 hours (20 of them at sea), or realistically one night in addition to the night on the boat. Total CO2 cost, ca. 110 kg. (You can also take various local trains down Denmark to reach Hamburg, but good luck with making the connections. Possible CO2 cost: 72 + 14 = 86 kg)

So, in short, I am describing a holiday trip with your family when you want to take it leisurely and take in the sights etc. It is not an option for a shorter visit, no matter how green you are.

The realistic alternative: flying to Europe
Thus my contention that going abroad for non-holiday purposes requires air travel. But that is not the end of it. I also had in my mind that it mattered how you travelled by air. If it is so that an airplane emits more on take-off than on cruising, does it not matter how many stopovers you have? And, as mentioned, the carbon calculators confirm that this is an issue. So, let us go to Paris once again, in two typical alternatives:

• Bergen-Amsterdam-Paris: One "long-haul" to Amsterdam, one "short-haul" to Paris: 291 kg.
• Bergen-Paris direct: 260 kg. 

      This may not be a major difference, about 11 per cent (although my averages may underestimate the extra cost of a separate flight). If, however, we consider the flight Bergen-Amsterdam (about 170 kg) the "unavoidable expense" of getting to the continent from Bergen, then distinction between the extra cost of a direct flight to a nearby destination like Paris becomes a bit clearer, 90 kg to 120. You also save travel time, of course, by some 1-2 hours, but connecting flights are more frequent, so it may be less convenient.
      However, the major difference is if you consider travelling the second half on the surface: Fly to Schipol, and then take the train to Paris. The train takes 3,5 hours, which means a time cost over connecting flights of perhaps only an hour or two, and the emissions - being mainly in France, which has pretty clean electricity - would be in the region of 3-4 kg. Now, this example may not be convincing compared to a direct flight when you already have "paid the price" of one take-off and landing. But it is fairly compelling for destinations that do not have direct flights from Bergen.

A pragmatic strategy for improving green-ness in European travel
So that would line up this strategy, when we go to the continent:
- Take a direct flight from Bergen to the closest possible airport, and then take the train the rest of the way, rather than messing with connecting flights. Thus, avoid going through Oslo or Copenhagen. Today, of course international and air travel is in flux, we do not know which airlines, flights or direct destinations will still exist when borders open and travel restrictions are lifted. London and Amsterdam are certain to be back, while Paris, Berlin or others may be less certain. But the general rule will probably still hold: If your final destination is within 5-6 hours train ride from the closest direct flight connection from Bergen (which will cover much of Northern Europe), then the green alternative is to avoid connecting flights, and use plane + train instead. If you have to use connecting flights, use as few and long hops as possible.
      That will come with a time cost, of course: quite likely it may add an extra overnight stay or two. But that is probably the overall consequence we must draw from travelling green: It will be less common to go from your home, shoot straight in to the meeting, then jump into a taxi and fly out immediately after. We must factor in a bit slower pace if we add trains or buses rather than air travel. But, in most cases travelling to Europe takes a full day anyway, and the extra time cost will when we get used to it, not feel so onerous as we might think.

***

That was it. Is this all obvious - bus and train is better than plane, always? Partly yes. But my calculations here have highlighted that isn't just being "good" or "bad", white or black. There are actual quantifiable figures involved. And the facts that taking the plane to Volda is justifiable, was a bit of a surprise to me, and in particular that driving a petrol car is almost just as "dirty" as flying a jet plane the same distance, unless, of course, you pile more passengers into your car (then you divide the car figure by number of passengers, of course). But even for Volda, you would need at least four people in the car to match a full airplane. Also, for me at least, the "plane + train" option in Europe may well be something I will try to follow when and if we ever get to practice that again. But each of you will of course have your own priorities and choices, and for those who live on the continent, this is probably just an academic exercise: You are well served with trains and coaches everywhere. Consider it a view from the peripheral north.

More Arabic and Arabist stuff

Only a year has gone by, and another blogpost already! - But again, mostly for people writing Arabic or transliteration of it on Apple products. I used to have a full website of this (now at org.uib.no/smi/ksv), but I have not had time to keep that updated, and anyway; with Unicode and standardization, the use for it has decreased. Most things now on the Mac do work with Arabic, with the latest versions we can write reasonably well Arabic even in Microsoft Word (although I still prefer Mac-only NisusWriter for this).

So, I can rather use this blog for the occasional burst of insight I may acquire. Perhaps it will even turn up in google searches, so somebody reads it. However, for today there are two items, one for the Mac, the other for iOS (iPads in particular).

First: iPad: More on writing Diacritics.

Below, you will find a blog I wrote last year (2018) about using the Smart keyboard (i.e. the plastic hardware) for writing transliteration on iPads. This has not changed, except that Apple now makes these hardware keyboards for all current iPads, not just the iPad Pro. However, it is still a very expensive solution, $159 for all - probably not worth it, if you do not want to have this keyboard for other reasons. But too much just for the transliteration.

However, there is a simpler solution on the iPad itself, no hardware required, which only costs $5, and that is worth it. I mentioned it in the blog last year, but only in passing, because it was not terribly convenient. But now it is (or maybe it always was, I just did not notice). It is the app Unicode Pad Pro (“with keyboards”, it calls itself). There is a free version, Unicode Pad Express, but that lacks the crucial element, so splurge on the $4.99 Pro version.

The method is like this:

In the app, select «Notepad», and “click” in the open text field to open the on-screen keyboard. This has five tabs on top, named ABC, I, II, III, IV, V. Those named II - V are empty, and you can fill them with your diacritics:


(1) Go to «Catalog» (the cabinet icon top left of keyboard), and in the Unicode catalogue locate the characters you need. It is probably easiest to find them in the option «Unicode Blocks». Long vowels (āīū) are in the block «Latin Extended-A», while emphatics, ṣḍṭẓ, are in Latin Extended Additional. ‘Ayn and hamza are in «Spacing modifier letters». For each character you want, press and hold, and you will see «Copy». Do that, return to notepad keyboard, select an empty slot and paste the character (press the slot and hold). Repeat until you have those you want in the order you want.

(2) When you have got your keyboard set up with the characters you need, exit the Unicode Pad app, and go to the Settings app, to General, and choose Keyboards. At the top, you see a «Keyboards» item, click the > beside it, and in the following choose «Add keyboard». You should then in the «Add keyboard» window, find «Third Party Keyboards», and Unicode Pad listed there. Select it to add to your installed keyboards. You should probably also choose «Give full access» to the keyboard.

(4) Now, your custom keyboard has been added to your iPad. Open any text program that allows typing. Clicking on the globe icon on the onscreen keyboard will cycle through your installed keyboards, among them now Unicode Pad; and you will see the same selection of ABC, and I … V layouts. Thus, you can type directly using this keyboard layout into your text editor, rather than having to open the Unicode Pad app and copy/paste from there. For your regular text, you can select the ABC panel, and so switch back and forth between ABC and your Diacritics keyboard as you type, all within the Unicode Pad keyboard.

Have your Mac read Arabic to you

The second topic is on the Mac. Many may have discovered that you can have the Mac read out to you any typed text, using a “computer voice”. Many apps (like Nisus and Apple's apps) have a “Speech” menu item that controls this, others like Microsoft Word and Safari let you do it through a command key. However, this is of less use for multilingual users, as e.g. French or Norwegian text is read out as if the words were English, i.e. virtually incomprehensible. Useful only if your text is written in the same language as the “speaker” has.

It turns out that you can control this, however. It is just put a place you probably would not think to look. Go to the System Preferences panel Accessibility, and select the Speech (Text to Speech) panel. There you will find at the top of the panel, the menu System voice, with perhaps several voices, in one or more languages. Select “Add”, this will show you a large variety of voices in many languages, including French, German and - this is the point - Arabic. There are three Arabic voices, Laila, Maged and Tarik, available for download. Download those you think you will use. Then, you can use this panel to choose which voice, and hence language, your typed text (or Safari page) will be read out in.

For applications like Word or Safari that do not have a “Speech” menu, you can set up a command keystroke (default is Option+Esc) which will read out any selected text.
          Unfortunately, it does not appear that the Mac (or the apps, they rely on the system for this) is able to link this choice of voice to any language setting in the app. In Word and Nisus and many others, you can define your text to be in “English (UK)” or “French” etc., for spell checking, but this has no effect on the reading language, you have to visit the Accessibility panel each time to choose your language. There is an option, however, to make Accessibility a menu bar item, so you can quickly go to this panel and change language for reading when you need to.

In the same panel, you can choose the speed of reading. For non-native users of Arabic, you may slow the speaker w-a-y down, so that you hear every syllable clearly - you can then even hear its choice of iʿrāb. Because, yes, the Arabic voice will read unvocalized text, and even conjugate the words in the sentence (presumably only if it makes any sense). It seems to do this even if you are offline, so the grammar analysis must be done on the Mac itself. I have not tested how accurate it is, but at least it gives a suggestion for how words should be vocalized and the sentence structured - as long, of course, as the words are in the Mac’s Arabic dictionary (which it must thus have somewhere), and are meaningful. All of this may be pointless to the native speaker, but for any learner of Arabic it may be a useful little aid, to get a suggestion of how a text could be read. And it is certainly a useful tool, in any language, for proof-reading or checking that you have typed an original text correctly, looking at the written original while you hear the typed text being read out to you.

Diacritics on iPad?

This blog has been quiet for years, and I am sure nobody follows it. But who knows. I just downloaded a new blog editor, and since I had a topic that might be relevant, I could test it out.

The issue is relevant for those who (a) use iPad or iPhone, and (b) write in strange languages, to wit: write in transcription (writing e.g. Arabic in Latin script, with precise indication of long and short a, etc.). Most who have this need, have organized this on the Mac (see https://org.uib.no/smi/ksv/Diacs.html , e.g.). But what about on the iPad (or iPhone)?

As for Arabic transcription, we normally need line above some vowels (āīū) and dots below some consonants (ṣḍḥṭẓ). In addition most use specific characters for the two sounds ‘ayn and hamza, ʿ and ʾ, although others only use inverted apostrophes, ‘ and ’. I have not found any solution to produce the special characters, except using a special program such as Unicode Pad Pro, which gives you access to all signs in all character sets, but in a separate program. 

But how about typing directly on the keyboard?

As most know, you have different on-screen keyboard layouts on the iPad as on the Mac: Go to System Settings: General: Keyboards, and choose “Add keyboard”. That gives you a list. The English keyboard has some diacritics, while other languages have fewer, in my Norwegian one, e.g. I have only ā and ū, but not ī. The standard English keyboard layout, however, has āīū as well as š and ž, but not dots under. How to get them? When you are typing with the keyboard up on the screen, press the a “key” (button?) and hold. You will see alternative forms, such as á and à, and also ā; drag the finger over to the variant you want and let up. 

Other layouts may have more, I discovered I had one called “Hinglish”, which added ḍ and ṣ to āīū, but not ṭ ḥ ẓ, so not sufficient for us. 

However, I just now splashed out on a physical keyboard (“Smart keyboard”) for my iPad Pro, and that in fact adds a lot more options. It is a small plastic keyboard which you attach your iPad to as a cover, but can unfold to reveal the keys, with Command and Option key. You still can select different keyboard layouts, but even with standard English, you do get dot under, but by typing two letters: first the dot-under key, then the letter it should be under. These “diacritic keys” are the following:

Type Option-a  [option key and a pressed together], let up and directly type e.g. i - gives you ī. Same for most vowels, option-a and e for ē, etc. 

Option-g gives dot under: option-g and d: ḍ, option-g and s: ṣ. 

Further: 

option-x line under following letter: ḏ
opion-c:  cedille under following; ş
option-w: dot under: ș
option-p: dot over: ṡ
option-r:  hacek: ǧ
option-y: yumucak: ğ

and many more remote ones, I have not identified all:
option-h: ogonek under following, but only under a few relevant characters: ų
option-j: “tilde” under following, ṵ
option-k: tilde over, ẽ
option-l: line through, ł
Shift-option-F: yoruba ɗ ɓ

Shift-option in many cases give the same as option, only thrown back on preceding instead of above following letter. 

This may not be relevant for many, since the physical keyboard is expensive in itself (I have not tested if third-party keyboards give the same result), and it only works with the expensive iPad Pro, which restricts it further. But if you are in relevant category of people, check it out. I was surprised to see so many options for typing diacritics on the iPad.

Mavericks and Mail, update

I said in my previous post two weeks ago that I would wait installing Mavericks on my main Macs until an update for the Mail problems arrived. Well, it came out last week (it even installed itself), and seems to have solved the smart mailboxes issues - at least the Unread count was correct on my MacBook Air, where I had Mavericks, and all mails there were visible and accounted for. So, I went ahead yesterday and installed Mavericks on my main office computer (an iMac, large screen, fast Fusion drive, excellent).

Alas, I had not tested it enough on the Air and perhaps not searched enough on the Internet. It was to be expected that Mail spent more time getting up to speed on my main machine, with 200.000 emails archived, it seems to be working still a day later. In at least one set of mailboxes, all mails appear to be empty: itemized and dated, but with no content. However, it is there (visible in TimeMachine), so the program is perhaps still working its way through, if not I can replace them from TimeMachine. Similiarly, the "count badge" for unread mails in a mail box also needs time to update smartly, but in a day or two it will, hopefully. Again, my advice from last time, to recreate the smart box definitions probably holds.

So, patience will prevail. But there are other issues that actually have been widely reported on the web and are clearly bugs and design flaws. There are temporary ways around them, however.

Sending group mail
The main error concerns group mail, that is sending mail to many people at the same time. This is the way it used to be done: In the Contacts (address book) app, you create either a "group" (a fixed group, by dragging cards individually into the group) or a "smart group" (dynamic, defined e.g. as "address cards that contain the word "Class2012""). Either way, you can email collectively to the group by beginning to type the first few letters of the group name in the To: (or, preferably, Blind Copy:) field in Mail. Then all items in Contacts starting with those letters show up in a drop-down menu, you click on the one you want, and it turns into the blue "cartouch" signet of a addressee, with a drop-down menu like "Expand Group" etc. Send, the mail goes to every member of the group.

In Mavericks Mail, you can do this for individuals, but there is also a new method, familiar to iPhone users: To the right of the To: field, there is a blue + icon, which gives access to the Contacts book and its groups. Click on a name, and it is inserted into To:. But that only works for individuals, not groups of any kind.

Now, what happens if you use the old method of typing a few letters? It varies. For fixed groups, it seems to work as it should: You type, select the correct item from the alternatives, and get the regular cartouch, both in the To: and Blind Copy: fields. But smart groups do not work like this. Sometimes you do get the list of choices, but no cartouch appears, nor does it expand into individual members when you ask for this (in Mail's settings). At other times you do not even get the list, and nothing happens when you type out the name in full. Apparently addressing smart groups do not work, which it clearly is supposed to.

But, in fact it appears it does. The feedback you get varies. Sometimes, you do get the list of choices, and when you select it, you will briefly see the name in angle brackets afterwards: "Class2012 <Class2012>" - although the name in brackets disappears once you Ok it. At other times it does not show you anything. But it appears in fact that the smart group function actually does work in both cases: Select the name if given a choice, or type it out if not (making sure you type the name precisely), press Send, and the mail will really go off to all members. At first I thought it only went in the first case (I wrote that ten minutes ago), but when I tried it out also just typing it out, with no indication from Mail it understood it as a group name, I heard back that the mail did go through.
     Why some smart groups give you the minimal feedback, and others not, I do not know. And I have not tested all my smart groups to see if they all actually work - it certainly is an error on Apple's part that they give no feedback, and of course that you have to rely on (precise) typing of a smart group name to select it. I would certainly suggest adding yourself as a recipient to groups you use, to confirm that the messages actually go out to the list
       A workaround can be used to give you peace of mind that it was sent as usual: Go to the menu "Window: Address Panel", find the smart group in question, click inside the list, then select all (Command-A) to select all names, and click on the To: (Copy:, Blind Copy:) buttons to address them individually. Not elegant - not if there are a lot of names - but will work for the moment.

Clearly, this is a bug, but it seems to me also a design error, in that the apparently preferred addressing method, the blue + icon, does not seem to have considered how to include addressing group mails.

There are some other minor bugs: While you can choose Bold or Italics for text you edit, it does not hold: underline works, size change and colour works, but bold and italics are forgotten.

Still hold back?
Now, many people do not use either smart mailboxes or smart groups, so they will not be affected by these problems. But they are not really esoteric, but fairly important and mainstream functions. Every new OS has bugs, that is to be expected, but it is peculiar that Apple released a version of the important Mail program, one that everyone uses, without checking these standard functions better. Personally, I think I will still hold back a few days before I upgrade my last Mac, just in case I find more problems, but these issues can be worked around, as I mentioned, for the moment. On the plus side, as these bugs affect fairly many users and is kind of another strike against Apple, it is to be hoped that they are aware of them and put some pride into quickly solving them. There is indeed reported to be a 10.9.1 on the way, but Apple released the first Mail fix in advance of that (untypically, they normally roll Mail app fixes into general upgrades), so we must hope these other issues will be quickly solved as well.

Mavericks and Mail

New term, new operating system, new problems in the Mac Manor.

Last week, I and I guess most Mac users who had turned on Software update warnings, were invited in large friendly letters to upgrade to the new system, MacOS 10.9 or Mavericks, for free! A novelty, as we remember, it was from System 7.1 in 1992 that Apple started to charge for system upgrades, so after twenty years a more friendly approach is back.

However, before I got round to accept the offer (you must set aside download time and re-organizing time etc. for a couple of hours), I received a warning from TidBits that it might be adviseable to hold off - there were significant problems with in particular how Mail works in Mavericks one should be aware of, before taking the jump.
<http://tidbits.com/article/14219>

It seemed however to affect particular cases, especially those who uses Mail mainly for gmail (Google Mail) and had followed Tidbits' Joe Kissel's earlier suggestions for gmail, which now are broken. I never did that, and gmail is only a secondary address for me, so I thought I should on the one hand be careful, on the other test things out to see how it worked. As it happens, I have a travel Mac, a MacBook Air (on which I write this), with a full Mail setup, but which I do not use daily. I decided therefore to install Mavericks there first, to see what happend.

The install itself is smooth, as you could expect - click, and go away and read a book or magazine for the hour or so the process takes (check in at beginning and end for giving various permission, but in the middle just let it work). I forgot the very useful advice to first download the installer, then quit and copy that installer app on a USB, before installing, that would have saved me the time to download it again for when my other Macs are to be upgraded (the installer deletes itself when Mavericks has been installed). Would have saved me 15 minutes or so, but never mind.

On restart, you have to give various passwords, AppleID and the Mac's admin ID and so on, as usual. I think I made a mistake, because I use different IDs for iCloud and iTunes, and probably gave the wrong one at one point, because I got a blank square covering some window and which I could not click away. But once I saw what was below (System Preferences), I could quit that and reopen, and then (in the menu) of the new iBook app log out of my iCloud account and into the iTunes one, where I have my books. Worked.

Mail issues
However, the main quest was for Mail problems. And yes, there were problems, and yes, they are sufficiently important that I am holding off a bit to see if Apple will fix them with a "10.9.1" bug fix before I install Mavericks on my main machines. If Apple does not, however, it may be possible to work around the issues.

First, a simple one. Each time I openend Mail (which was a lot, in the testing), it kept bugging me for my iCloud password for getting @me.com mail (like everyone, I have such a passive account). It would not remember the password I put into Mail's Accounts pane. The answer was quick, and cited in the TidBits article above: Open System Preferences, the iCloud pane, and then first uncheck the Mail item, and then immediately check it again - that forces Mail to re-configure itself, and you will see that the password item has disappeared from Mail's own iCloud account setting - it now takes it from the general settings, and remembers it.

But then the major one. Joe Kissel's problem was, as mentioned, related to gmail, and had something to do with gmail's special "All Mail" mailbox. I had never messed with that, and could see no problems of the type he mentioned. Probably this, the major point of his article, only or mostly concern those who use Mail mainly for gmail.

However, the second point he briefly mentions, and which has been widely reported on the web, was a major stumbling block for me: Smart mailboxes, which were and still are mainly broken in Mavericks Mail. Smart mailboxes have got several new features, and are possibly built up differently from before, I don't know. However, to explain, this is how I read mail, and why smart boxes are important for my workflow:

I use Mail as a main mail archive, set up with dozens and dozens of mailboxes: for persons, for topics, for mailing lists, for countries or whatever. Anything that I can identity (mainly from regular senders) gets automatically filtered to its mailbox immediately on arrival, whether it is my students, from the department, or newsletters or mailing lists - everything disappears from the In box before I see it, but is of course still listed as "unread" until I read it. The In box only includes mail from irregular correspondents, or those colleagues I want to read first - I always first read through the Inbox, dump spam and answer requests. Then, I have set up a smart mailbox that collects all unread mail that is (a) not in the Inbox (b) not in the high-volume mailing lists, which I read at the end when I have time. But this "Unread" smart mailbox is thus essential, it includes lots of important mail from regular senders.

What happened in Mavericks, is that the Unread list gave a surprisingly high number of items on its badge: about 200, while it should actually be 10-15 since my last mail check. But when I opened the box, it was empty. The menu line said: "0 messages. 196 unread". Now, I could see down in the list of "real" mailboxes that there were a number of unreads down there, but the smart box did not list them. I restarted a few times and tried a few others things, like creating new smart mailboxes afresh, but with basically the same result, all apparently empty. I saw on the net many people who had tried various strange suggestions to no avail, so it was and is a real bug.

Finally, I found a solution that was supposed to work (in the same TidBits article; always check in Tidbits): "Kirk McElhearn found all his smart mailboxes empty, and was able to correct the problem by quitting Mail, deleting the three files beginning with “Envelope Index” in ~/Library/Mail/V2/MailData, reopening Mail, and letting it reindex his messages." I did this,(*) and lo and behold, something happened: Mail put up "Welcome to Mail! For this new version, we need to process your mail to the present version of Mail" [or words to that effect, from my memory]. This is well and good, it is supposed to do this, but it never did when I first started Mail in Mavericks. So here was a possible bug, Mail had skipped its own upgrade process. Anyway, it went to work on my 100.000+ messages, and spent the next fifteen minutes on that. And did it work?

Yes! Or at first appearances anyway. The message count in Unread was down to about 25-30, a bit too many (it included a number of old messages from gmail's Unwanted mailbox), but they were all present and viewable in the mailbox. So, I could just walk through them as ususal and see the badge count drop by one for each. Down to the last. But that one did not go away. I dumped it manually, and it did go away from the mailbox, but the counter remained at 1. The same for the new boxes I had created just before, the all were stuck with one, invisible, unread mail. Now, I found a way to get rid of that minor issue: I went to Edit smart mailbox, took away the rule "Message Is Unread", and then added it again, and then the invisible 1 disappeared. Everything OK. Until the next actually new unread message came in - but not in the Unread box. This time, neither badge nor content shifted, the smart box had simply stopped working. Only the second (old trial) box, where I had not edited the rules, worked (with the invisible 1 in place), all others were broken again. But that one worked - or did it? It included several unread messages that were in the Inbox - which should be excluded. So, it seems to have found all unread messages, i.e. just the first line on the smart box definition, not the ones following.

Last ditch effort: To recreate a smart box afresh, now completely under a properly updated Mail. That kind of works: The invisible 1 was back, but otherwise it seemed to pick up actually unread messages in the mailboxes it should (although some have got strange names, like "Archive" and "Important", apparently from gmail which Mail formerly ignored) and only those. This morning, after one night, the invisible 1 even went away again, but now in the afternoon, it lists 3 after I have read everything. So the badge count is off, apparently haphazardly so, but I can live with that, I guess. I will keep watching for a while, and see what happens.

But the moral appears to be: Make sure Mail actually shows a Welcome to you message and spends time updating your mailboxes (you will notice, a lengthy process), and then re-create any smart mailbox you have had in the old version. For the monent, this seems to work, but there may be other bugs that could raise their heads (but I am very happy that Cmd-Option-T for "Move to [last used Mailbox] again" once again works, saves me a lot of menu trips).

(*) How to open your Library folder? Improvement: It is still hidden by default, but now you can set this to display in an option in Finder's: View: Show View Options menu.

Netflix and the new Macs

I don’t normally write about new stuff here, everyone else does, but yesterday as I was looking for baseball news (go Giants!) on my fairly recently acquired Apple TV, I noticed a new icon there, “Apple Events”. Wondering what that was, I was taken to the presentation Apple chief Tim Cook made only three hours earlier. So I got the early not-news of iPad mini. However, also all the other news, Apple presented new models across the board: MacBook Retina, mini, iMac. And something stood out: we are facing a shift in how Macs are equipped. One change is from hard disks to flash drives in most new models, not necessarily good for the checkbook, as flash of more than the minimal sizes are very expensive (but the new combo “Fusion” drive looked enticing, pricing to be seen).

Even more clear is that DVD is out, no new model has a DVD drive (thus the Air’s configuration has become the general pattern), and it was derided as “you old-fashioned people” at the presentation. The signs have been there, Mac OS has not been available on DVD for a while and hardly any programs are distributed this way. And in entertainment, streaming has replaced DVD.

Or has it? Is streaming the way to watch movies and TV shows outside the TV set? This links to the other big news in my country last week, which is related to the same topic. OK, what follows is not so much about the Mac, but about streaming in general, and is also a posting strictly about Norwegian matters. I might have written it in Norwegian, but let me continue in English since I started that way - the rest of you can see it as a case study from a small European country.

The advent of streaming TV

The big news, then, was that Netflix has arrived to Norway, and that HBO is due to follow in a few days as both streaming and a TV channel. This (Netflix) will change the way we watch TV, it was announced, now we will get all new series instantly online (legally) and not watch them on TV any longer. Of course, this is not true, and was a confusion of the two services: HBO promises immediate transmission of its new series to us, but that is not Netflix’s business. Netflix was originally an innovative video rental service, which switched to DVD rentals and then to streaming. Thus they do not distribute new series or films, only once they have arrived on DVDs, that is half a year or longer after original release.

I know Netflix from when I lived in the States, and loved it then for its rapid DVD delivery, and apparently bottomless catalogue; they seemed to have virtually every film or TV series ever put on DVD. They were just introducing streaming when I was there (2009), and that did not impress me so much; quality was low and selection meager. Netflix in Norway is streaming only, and is not the first such streaming service we have here, but arguably the most convenient to use, with a fixed monthly rate giving unlimited access to their catalogue. Of interest to us, Netflix is available on Apple TV, iPhone/Pad as well as on the web.

As for HBO, it does promise immediate release of new series, but it is unlikely to make any dramatic change in our viewing patterns, and in particular not for our main TV service NRK, as was darkly predicted. It will, to go into the national details, rather be a problem for Canal+ (now CMore) which has been the channel for new HBO shows. Currently, the pay channel Canal+ has sent these shows about three-four months after their US release, then NRK has re-sent the major HBO shows to a much wider public half a year or so or after that again. That HBO enters the chain as a pay channel will thus not make any significant difference for NRK, but the question is whether all these top shows will now disappear from CMore. Possibly for this reason, CMore has precipitated the new seasons for most of their HBO shows this autumn to send “just a week or two after the US”, so that most of them this fall will be finished by the time HBO itself opens later this month. But what of next year? Their “Series” channel will be very thin without the HBO shows, but is it possible to have a three-stage dissemination showing them first on the new HBO channel for their customers immediately after US release, then on CMore for their customers a month later, and then on NRK half a year after that? To be seen.

As will the form of distribution, HBO will both stream and become “a TV channel”. But how will that channel be distributed? As part of one of the three or four main “film packages”, CanalDigital (satellite + cable), Viasat (satellite) or Get or Altibox (cable / IPTV)? If HBO joins the first of these, CDigital which has most HBO shows already, it will mean hardly any change whatsoever from today, except perhaps a small price bump and slightly faster transmission after US release (perhaps cutting down on illegal file transfer, but little else), if one of the other it simply means the HBO shows transfer from one major channel package to another. Well, the answer will be revealed in the near future, but in neither case will this “change the way we watch television”.

Streaming movies: A comparative test

That clearly relates to streaming to our Macs, PCs or tablets instead of watching traditional TV channels. But how useful is that really today? We already have several such Video On Demand (streaming) services in Norway, and on the occasion of Netflix’s arrival, I decided to make a survey of them to compare with the Netflix offer. There are about ten of them, offering mainly movies, typically for ca. 30–45 kr per film (one service, Voddler has like Netflix a fixed subscription, but adds a per-film charge in addition). They all “suggest movies you may like”, but to be really a new experience, what I would want to is look up a movie I know about, and find it on their service. So, I set up a list of 26 haphazard movie titles from my shelves, mostly from 2005–2010, from LOTR and other Oscar winners to current comedies, and searched for them on each of the services. In addition, I looked for films by two famous directors, David Lynch and Woody Allen, and in those services that showed categories, I took their suggestions for “Science Fiction” to see how many titles they offered (the latter is of course less useful, as each service may differ in how widely they define this category. Unfortunately, few of the services tell how many titles they have in total, except “thousands”, two say specifically 3,000+ titles. In comparison, Netflix's original DVD collection had 35,000 film titles).

The ten are: Voddler, Comoyo , film2Home, Lovefilm , SFAnytime , Filmarkivet , CDON.com, Homebase, CanalDigital, and ViaPlay.
What I expected was, first that only a few of my sample titles would be available for streaming, and that it would be the same titles for all ten, since the limitation was distribution rights from Hollywood, which would be limited country by country. In fact, I was surprised. The first assumption held true, out of my 26 total no service had more than 9, and the majority only 4–5. But the second was not true, every service had a different set of titles, so that in total only six of the 26 titles were completely unavailable. I do not know the reasons for this, but it suggests to me that the limitation is not Hollywood (distribution), but investment from the services; evidently it costs money to build up a library for distribution.

As for my directors, no service had any film at all by Lynch (whose last major movie was Mulholland Drive back in 2001), while the score for Woody Allen was surprisingly good: almost every service had a majority of the movies he has made after 1994 (from Bullets over Broadway), but absolutely zilch for his earlier great movies. But of the new stuff, at least two (Comoyo, SFAnytime) had 17 out of a possible 18 titles post 1994. SciFi? Voddler claims 118 titles, Viaplay 59, Comoyo 55. None a terribly great number, when you think of how many movies can fall in that category. In total, which service “won”? Well, it is not fair, since my title list was so haphazard, but in this very non-scientific test, film2home and Viaplay scored best with 9 and 8 of my 26 titles, the rest stretching from 1 (Filmarkivet) to 6 (Voddler, Comoyo, SF Anytime, CDigital).

So how does Netflix stand up to this? As expected, pretty much the same, not dramatically better but not worse either. Of the 26, it scored 8 titles, which is good, less good on Woody Allen (6 of 18). But they had one old David Lynch. So, Netflix is certainly more convenient, in particular if you watch more than a couple of streamed movies a month (and even more if you have Apple TV), but it is not really a game changer for the great public. I must add, however, that Netflix also streams TV shows (those that have appeared on DVD), and I have found quite a few of my favourites that never came to any TV channel here, so that may be a noticeable advantage over the other services, only some of which offer TV shows.

Of particular interest - and gave me a great surprise - was the question if it is “Norway” that is bad - whether it would have been dramatically different if we had had access to US Netflix. I was able to make a quick comparison with Netflix US, and was astounded to find that of my 26 titles (all Hollywood movies), only 3 were available in Netflix US, and none of the later Allens (but three older ones); far more limited than the Norwegian store. I can hardly believe it, but unless I did something quite wrong, there is no reason to chase after Netflix US rather than the Norwegian version, in terms of breadth of catalogue.

So what is good? (Fanfare:) iTunes. That is, iTunes US - which I was incorrectly guided to before I logged in with my account. They had almost all of my 26 sample titles for rent or purchase (23 / 26), and most of the classic Allens (but only half of the new ones). But, unfortunately, once I had logged in with my Norway account, it all went away: only 4 of the 23 remained and a single Allen. Norwegian iTunes is basically of limited value. But what this shows is that all of my 1990s–2000s sample films do exist in streaming format somewhere (and not just illegally), it is just a matter of distribution and commercial choice to let us access them.

How does this link to the new Macs? I think this is clear: Streaming is not yet mature enough in width of offering to let us drop DVD. If I want to see a particular film or one particular show on my Mac, I will in most cases have to get it on DVD and play it from there. It is possible that streaming will, like Spotify for music, extend their choice so widely that I can get any current or non-current movie from a streaming service, even in a small market like Norway, but we are not there yet. So, buy the external DVD drive, if this is something you want to use your Mac for.

But I will keep my Netflix subscription.