The lead image in the traction engine article in various languages

One thing I like to do from time to time is compare the images different languages of Wikipedia are using for the lead image of equivalent articles. While in a lot of cases the pattern is; whatever the english wikipedia was using at the time the article was created that isn’t always the case. And for traction engine it mostly isn’t. One complication is that for many languages it seems the article would be better translated as “road going steam engines” but the results are still interesting.

First off we start with English and Spanish (which appears to be a translation of english). This is my image from the Great Dorset Steam Fair. I went with it as lead image worth as it is a decently clear image of a pretty standard traction engine (not a ploughing engine or showman’s engine).It is pretty isolated and faces left (and thus into the article) with a three quarter view. Its not ideal as the background is busier than I would like and the crop is a bit on the square side since I removed an arena marshal. It was taken during a concerted attempt at great dorset steam fair to take the best traction engine pic possible. Unfortunately they drive round the arena anti-clockwise which makes getting left facing images rather tricky (the engine in the pic is cutting across the arena). This was driven by the previous lead image being rather dark and blury although fairly understandable for a product of 2005. Technology and wikipedians have marched on since then.

Second is Danish. This is a portable engine not a traction engine. It appears that danish is one of those languages where “Lokomobil” refers to any mobile non-railbound steam engine (or at least that is what the article does). The image itself is a black and white period image showing a portable engine at work. There are some who argurage that that is the best way to illustrate the subject. Personally I argue that humans see in colour and thus its best to lead with a colour image if one is available. The quality of the image isn’t the best (I suspect its a second generation copy) but the scene is well laid out and 1895 is very much the time period where these things were in use.

Third is German. It is a traction engine this time. A ploughing engine in fact. Not the greatest image though. Low res, not the sharpest, blown highlights and partially obstructed with a fence in front. Points for facing left though. In practice this wouldn’t have been that untypical for 2005. Which is when the image was added to the article and it has been there since. The machine is apparently in Germany which may help explain the choice of image.

Fourth we have French. “Locomobile automotrice” apparently translates as “Self-propelled locomotive” which would explain why the lead image is a steamroller (although it was described as a traction engine on commons until I fixed it). I think it is a film scan and the quality is certainly OK for 2006 when it was uploaded. However the rear view is an unusual choice for a lead image (although it does show a lot of the working parts better) and facing right it faces out of the article. Like the German one it could probably do with updating.

Fifth is Indonesian . It is a traction engine apparently at a show in Stuttgart. Its somewhat blurred due to camera movement (1/8 of second exposure) And faces right out of the page. Other than that its fairly well isolated, the view isn’t obstructed and the background isn’t overly messy. What I don’t understand is why they swapped out the previous photo of a very similar engine. Yes the front wheel is slightly obstructed but its less blurry.

Sixth is Cornish. The composition is decent with a nice action shot and the engine (which is a traction engine) facing into the article. However its rather low res and the back of the engine blends with the background. My guess is the author decided they didn’t like the then lead in the english language article and chose this from what they could find on commons.

Seventh is Dutch. This is a portable engine. Since its labelled “Locomobiel” the person who added it may not have known the difference. It is however a very good image of a portable engine. Its sharp and well isolated. Any improvement is going to be mostly stylistic. For example a preference for engines that are running rather than the funnel being in the stowed position.

Eighth is Japanese. It does show a traction engine (a showman’s engine). Its facing left into the page. However beyond that I’m not a fan. The obstruction by the fence is fairly minor but I feel its overwhelmed by the structure around it. The colours also look like they have been pushed a bit hard and may have been over-sharpened. A combination of these factors and the existance of better options was why I removed it from the English wikipedia. The engine is in Japan which may explain the choice here.

Finally we have Finnish . If you wanted an action shot of road locomotives then you couldn’t do much better. Yes it faces out of the page (great dorset steam fair going clockwise strikes again) but its a well composed photo showing some serious heavy haulage. Conventionally it might be better if the engines were a bit lower down to follow the rule of thirds but I think it still works. Personaly I prefer a single isolated example for the lead but in shorter articles that may mean losing the ability to show the range of designs.

Wikidata uses a perfectly decent film scan image although not perhaps as isolated from the background as it could be.

I’m not arguing that all the articles should be changed to use the english image. Asside from the obvious conflict of interest different language communities will have different priorities and I don’t speak any of the languages in question so there may well be things that I’m missing.

Posted in wikipedia photography | Tagged , | Leave a comment

How bad is Chia going to get?

For those less interested in hard drive prices Chia is a newish cryptocurrency that is mined on hard drives. Similar to the old burst coin.

Given what traditional cryptocurrency did to GPUs the question is what is it going to do to hard drive prices. Exact numbers are impossible since they depend on the ongoing price of Chia and behaviour of the miners.

Headline figure is that at current prices Chia can be expected to use about 100 Exabytes of storage over the next year or about 10% of annual hard drive production. At current prices its going to get pretty bad. I can’t predict how bad but I really doubt there is anywhere close to 10% slack in the hard drive production market.

That figure does ignore the cost of electricity which would take us to about 95 Exabytes and more significantly ignores the fact that that much buying would case the price of storage to rise (but I have no way to calculate that).

The calculation

The calculation is based on the fact that chia mints 9216 new coins per day (in pairs for some reason). Current price is ~$600 per coin. Resulting in $5529600 total per day or $2.019 billion per year. A year seems a reasonable upper limit for miners to get a return on their initial investment (hard drives don’t hold their second hand value well). Prior to Chia storage was around the $20 per TB mark and at that price $2.019 billion pays for around 100 Exabytes.

Electricity for storage appears to be in the region of 10kWh per TB per year at a price of $0.12 per kWh that bumps the price of storage up to $21.2 per TB which reduces the amount to 95 Exabytes. The impact of the cost of motherboards, NVMe drives to write to the drives and land to store the whole thing is beyond my ability to estimate.

If instead of looking for return of initial costs in under 1 year miners focus purely on electricity costs then Chia maxes out around 1.6 Zettabyte (higher depending on the price of electricity and the amount stolen). Global annual hard drive production is only around 1 Zettabyte.

Current growth rate

According to this chart Chia is currently taking about a day to add 500 Pebibyte. In units we are more use to that amounts to 562950 terabytes or nearly two wikimedia commons. At that rate we will be reaching the 100 Exabyte mark in around 180 days.

How much it needs to fall not to be a problem?

Using the above calculation $64 per coin would result in a 10 Exabyte pool or about 1% of global hard drive production which might not be too bad.

Posted in Uncategorized | Leave a comment

Handbreak now supports webm

Handbrake has long been the best option for those that want to transcode files to a commons friendly format while not messing around with the command line. Unfortunately it output files in the Matroska container (.mkv) and the easiest way to get them into a format that commons supported involved messing around in the command line with ffmpeg. However with version 1.3.0 Handbrake now supports webm directly simplifying transcoding.

Posted in Uncategorized | Leave a comment

An image for Pinchbeck from commons

I’ve been aware of Pinchbeck for some years after seeing an example of a Pinchbeck hairband at Bristol Museum & Art Gallery. The Wikipedia article on the alloy has needed a pic since 2005. Unfortunately my pictures from Bristol sucked due to missing focus and simply not holding the camera still (in my defence I was running on about 2 hours sleep and trying to visit all the museums in Bristol).

Getting a photo of Pinchbeck suffers from a problem of being sure what you are photographing is actual Pinchbeck. The term seems to be used to label 3 different things. You have true kvlt pinchbeck from the original Pinchbeck family, approximately contemporary imitations and then a random collection of vintage brass.

To be on the safe side I wanted a photo of the first option. This rules out simply buying some since the well provenanced stuff is expensive and for the rest I don’t have the expertise to dig through. Museum materials seem a better option but being somewhat reflective it tends to photograph poorly under museum light. Birmingham museums has some but the lighting in that section isn’t the best (although that can be countered to a fair extent).

Finally I tried a commons search. I don’t usually do this because a lot of the time you don’t get anything and the quality can be rather mixed (the best commons search engine remains the english wikipedia). On this occasion though After a lot of photos relating to the pumping station and the village in Lincolnshire I found this clock made by the son of the original Christopher Pinchbeck. Definitely true kvlt Pinchbeck

Posted in Uncategorized | Leave a comment

Wikipedia, right, but on the blockchain round 2

Well its more a re-launch of Everipedia which has been around for a while. Everipedia is essentially an attempt to make money by starting with a mirror of wikipedia and then going forward with no notability standards. Income attempts are by charging people for article writing services and adverts.

So far success seems to have been fairly limited. Low alexa rank and a fairly low rate of linking from reddit a fair number of which focuses on stuff of interest to reddit’s nastier sub communities. It also has the issues you would expect of something with no notability standards and has become a popular platform for the likes of 4chan to libel people

Still I’ve seen worse.

So where does the block-chain come in? Mostly in the form of what appears to be (or what is at leasdt marketed as) a relaunch which uses much the same language as Lunyr. Decentralised, censorship resistant and on the block-chain. In practice this seems to boil down to them thinking about hosting on the interplanetary file system (so Lunyr again).

One trick they have beaten Lunyr on is hiring wikipedia co-founder Larry Sanger who’s name is apparently enough to get your press release read. Quite how a project that claims to be removing gatekeepers is meant to work with a guy who’s previous project (Citizendium) was effectively a nested pile of gatekeepers I’m not sure.

What this is all in aid of is an initial coin offering. Called IQ, the coin’s nominal use case appears to be buying the right to challenge edits on the wiki (although things are at the unpublished draft stage). How they plan to make that something people would want without completely breaking the wiki system is unclear but I personally doubt they will ever have enough traffic for it to matter. If they get their coin out before the ICO bubble bursts they might raise some cash (after all they have the advantage of a product that actually sort of exists) otherwise not so much.

Everipedia is a project that has been trying to compete with wikipedia for 2 years now even as someone who keeps an eye on wikipedia competitors its barely on my radar. I don’t see anything in the relaunch that would change that.

As for Lunyr they continue to update their advertising system but since everything is still a behind a closed alpha its hard to say what’s going on. They are though apparently having issues with spotting bugs in their software due to CryptoKitties messing up the Ethereum blockchain. They’ve also apparently been removed from the Liqui cryptocurrency exchange but I really doubt that will matter to anyone beyond the unfortunate compulsive gamblers who trade the thing.

Posted in Uncategorized | Leave a comment

Monkey selfies and Technollamas

The monkey selfie is back thanks to a This American Life program. While it mostly deals with Slater vs PETA and gets that right its coverage of wikipedia’s role is more questionable. Techdirt has the details of that:

From the copyright nerd POV the most interesting fallout is Technollama’s attempt to do an analysis of the case under UK law:

While I broadly agree with their analysis (although I think they underestimate the differences between civil and common law copyright) a lot rests on the statement “If we believe Slater’s own telling of the story”. The reality is Slater’s telling of the story has been inconsistent. The initial version had the monkey picking up the camera and the whole thing being unplanned. There are reasons to be sceptical of the camera on a tripod claim. In particular one of the shots shows Slater resting his left hand on a tripod. I don’t exactly travel light in photography terms but I don’t carry more than one tripod unless I have a car with me (and even then the second tripod is a mini one). Other photos in the series were taken at different heights which again suggests a tripod wasn’t used. Technollama also argues for selected the lens aperture. Its possible. With wide angle lenses its hard to judge the depth of field well enough to tell. However the exposure (checked the lighting) jumps around a fair bit between pics depending on how much of the money is in shot (most obvious by looking how light the leaves are in the background). A fairly clear sign of the camera controlling the exposure (a human would be more likely to under expose a touch to try and avoid blowing the highlights before trying to bring the shadow detail out in post).

Post brings us to Slater’s actions after the picture was taken. My feeling is that this is Slater’s strongest case. None of the images are at the camera’s native resolution or even the same ratio as the camera’s native resolution suggesting some rotation and cropping. Its impossible to say if the colour balance has been changed. Does rotation and cropping qualify for copyright? Perhaps although the UK’s Intellectual Property Office ,“it seems unlikely that what is merely a retouched, digitised image of an older work can be considered as ‘original’”. Does cropping and rotating count as merely retouching the older monkey produced image? How would the courts rule? I don’t think there is any direct case-law yet.

In the meantime we are getting a bunch of emails to OTRS blaming wikipedia for Mr Slater’s issues and financial position. This is I’d argue somewhat unfair. The raising of the the issue of the image’s copyright status started with techdirt not us. More broadly the problem is due to the changing nature of the wildlife photography market. We now live in a world where you have a bunch of people who can afford high end camera gear and actively enjoy taking it to strange places and taking pictures of wildlife with it it. While these people have always existed in the past it wasn’t easy for them to offer their images for sale. Now it is. Being in the right place with a decent camera and the ability and willingness to sell you photos isn’t worth what it once was.

Posted in Uncategorized | 3 Comments

Wikipedia right, but on the blockchain

Well Lunyr, the project in question, is more google knol than wikipedia but they aren’t going to say that in their adverts.

Their system is essentially the knol model with writers getting a cut of the advertising revenue although they have also thrown in a review step. This is the short version; the full version is rather messy.

So far so boring. They do have ~$15 million backing it though with the exact amount depending on Cryptocurrency fluctuations.

Okey some background. The latestish thing in cryptocurrency (think bitcoin and various related things) is Initial Coin offerings. The idea is you buy a cryptocurrency that the startup you buy it from will use for some aspect of their business giving it long term value. In practice at lot of these are scams, ponzi schemes or just pure hot air. Lunyr where the idea is that Lun (the name of their coin) is what you use to buy adverts on the site is one of the more coherent ones.

Lunyr is still in private alpha so much of this post is based on screenshots, Lunyr’s own claims and various comments by people with access.

Broadly I don’t see it getting very far. If google couldn’t get knol to work I don’t see anyone else doing so. The reviewing step I suspect will be less helpful than they hope. After all wikipedia surpassed Nupedia precisely because it didn’t have a reviewing step. Citizendium tried something similar and again it didn’t work to well.

Lunyr tries to get around the problem by rewarding both writers and reviewers. The problem is that the payments are unlikely to be high enough to make either worthwhile. Writing a comprehensive wikipedia article can easily take 100+ hours and reviewing such an article to a high standard (in particular checking the sources ) can again be multiple hours of work (even if the sources are fairly accessible). For anyone wanting to write an article for reasons other than money well wikipedia and a bunch of web-hosts already exist. Equally advertisers already have plenty of places that advertise on the web most of which accept $£€ etc rather than making you go through an exchange to buy something you have no other use for.

If they are somehow successful in earning money then system gaming becomes not just a problem but probably the single biggest factor in the system. Consider all the scams used to mess with impact factor in scientific journals then turn them up to 11 in a world where multiple identities are trivial. Registering a bunch of reviewers and using them to reject any article who’s author hasn’t paid you off is simply the most obvious attack.

On the presentation side their layout appears to be the one critiqued here which basically says it focuses on being pretty over being useful. That said with the greater divide between writer and reader and the importance of looking pretty on mobile platforms this may not be much of a problem.

In terms of the cash they have $15 million. This may seem like a lot but Encyclopedia of Life had $50 million and a far more solid support base and still doesn’t see much use.

Full release is probably meant to be in the fourth quarter of 2018 although their description of what that will entire is pretty vague. As in literally “peer review system 2.0”. Meanwhile the link their whitepaper (the thing Initial Coin Offerings mostly write since its part of the ritual) is now dead. They also currently host on InterPlanetary File System which might be a problem if they gain any traction (although I don’t think they will.

Lun is currently declining in value on the exchanges but given how thinly its traded I wouldn’t read much into that. Fundamentally though there is no reason why Lunyr couldn’t just decide to sell adverts for $,£,€, PLEX, CS:GO knives or even bitcoin. Or take they could just what’s left of the money and run. Initial coin offerings don’t really offer any buyer protection.

Posted in Uncategorized | Leave a comment

Canon EOS 5D Mark III donation

My old 5DIII is now in the hands of Wikimedia UK meaning that they now have a full frame camera. While its a decent camera there are some catches. It has been through about 42K shutter actuations or a bit under a third of its expected life. It has had zero maintenance during that period (no sensor cleaning for example) and not exactly spent its time indoors as a safe studio camera. Its been as far north as Dunnet Head, as far south as Sandown, as far west as Delabole, and as far east as Norwich.

Attitude range is rather more limited. While it has been at sea level I’m not sure its been much above the summit of Arthur’s seat.

That said with the low light performance you can only get from full frame, good autofocus and 22.3 mega-pixels it still have as a lot of use left in it. The real limit it is going to hit is lenses with WMUK currently only having 2 lens that work with it both at 50mm. While there are “50mm only” adherents it is rather photography’s equivalent of iron-man mode.

Posted in Uncategorized | 2 Comments

The Canon EF 11-24mm f/4 for Wikipedians

Its a £2700 lens. At that price I suspect anyone buying it can come to their own conclusions. Still on a full frame camera it is an extremely useful lens. The width makes it great for urban architecture, larger items in museums, and interiors in general. The short minimum focus distance makes it great for objects in cases and the lens’s sharpness makes it viable to crop the resulting images.

Obviously if you want to shoot longer than 24mm then you need another lens but for wide angle work the lens is excellent.

Downsides. Its a £2700 lens. You could buy quite a lot of other gear for that. The Sigma 12-24mm f/4 is about £1000 cheaper and nearly as sharp at the wide end. If you are shooting on a crop sensor then the 10-18mm is under £300 so unless you really really need the sharpness for some reason I wouldn’t go near this lens for a crop system. On top of that its big and its heavy. Not something I have an issue with but for anyone more weight conscious (but then why shoot full frame?) it may present a problem. The f/4 speed may be less than idea for indoor work but thats becoming less and less of a problem as camera low light abilities improve.

Overall a very useful bit of kit but also really rather on the expensive side.

Posted in Uncategorized | Leave a comment

It is open acess week

Open access week runs from the 24th of October to the 30th. That is why the featured picture of the day on the English wikipedia comes from PLOS ONE.

Open access material appears to have fallen of the wikimedia radar a bit in the last year perhaps because a lot of it is rather hard to use. Probably the biggest issue is that a lot of papers cover a very narrow area within which wikipedia doesn’t even have a general article. Some of this is simply because wikipedia is weak in certain areas but others look suspiciously like publication padding. There are various strategies to deal with this but most of them run into the next problem.

The is a lot of junk in the databases. Sure if you stick to PLOS ONE most papers will meet a certain minimum standard. Beyond that things get interesting. Anyone can start an open access journal and while there are many decent attempts to do so there are also a lot of scammers and cranks out there. Scammers target the pay to publish model (which a lot of open access follows) by setting up a legitimate looking website and then accepting everything sent to them where the author pays the relevant fee. Cranks just want somewhere to publish their cold fusion and free energy papers that looks legitimate. Checking against Beall’s list provides a degree of protection but well, it is big, new journals pop up all the time and it doesn’t list the journal responsible for this paper arguing that the old Venera images show lifeforms on the surface of Venus.

So if want to add open access content to wikipedia how to go about it. I’m sure the are various methods but this is mine. Firstly I’ll identify journals that publish a lot of papers that can be used for sourcing wikipedia articles. Archaeology, astronomy and history are generally good choices here but there are others. I’ll search for these through DOJA Secondly I’ll see if I understand them. The European Geosciences Union publishes some great journals and the ones covering space science are a good source of things to cite but you need some serious physics to understand a lot of them. Thirdly I’ll make a judgement as to their reliability. I tend to do this by checking who they are affiliated with (if anyone) and if I trust them. I’ll also check them against Beall’s List and just by poking around the site a bit to see what they have published before. After that its just a matter of checking them every few months to see if they have published anything new of interest. Examples of journals I do this for are Internet archaeology and the Journal of Lithic Studies.

Another approach is to get personal recommendations from people active in the field. An example of this would be Polar Research.


Posted in Uncategorized | Leave a comment