Met Office cuts off Linux users with new weather widgets
Linux users face increased inconvenience getting a weather forecast from March onwards when the Met Office will withdraw its web-based weather gadgets and replace them with desktop widgets – for Windows and Mac only. Previously the Met Office's Firefox and iGoogle weather gadgets allowed anyone with internet access to check the …
I don't see the Met Office bending over to help the Amiga, Atari ST or ZX Spectrum platforms for that matter. Development costs time and money, so to recoup that they need to "pander" to the most common platforms to see a reasonable ROI.
The Met Office should dump Flash, and use a platform-neutral technology instead.
"Development costs time and money, so to recoup that they need to "pander" to the most common platforms to see a reasonable ROI."
Picking the wrong technology and doggedly ploughing on with development also costs time and money.
One would hope the Met Office would have the sense to implement one backend which can serve any number of different front ends over a simple JSON or XML request. Then they can throw the APIs open so that platforms which are not officially supported can still receive data through 3rd party apps. e.g. someone could write a GNOME shell extension which called the same service. Done properly, the Met Office probably even make some money off it.
Publishing an RSS feed or similar costs time and money, but not much.
They charge money for their an XML feed (which is basically giving you an FTP read access to a DAT file), that is why they won't have a XML/JSON api.
Agree entirely.
As a Linux only user I might be devastated by this news except I don't use Met Office widget for Firefox anyway.
"but that involves at least one more click than usual" - er ? Bookmarks anyone ?
As for iPlayer - simples use get_iplayer instead
You want a site delivering Met Office weather forecasts that isn't tied to a particular OS because of its uses of non-standard tech? Try http://www.bbc.co.uk/weather/.
You want something to embed on your desktop because you can? Congratulations. You are an insufferable geek and will just have to live with the fact that the rest of the world does not exist to pander to your taste in consumer electronics. Frankly I think it is silly that they've developed widgets for Windows, let alone any other platform. Can't they just spend the money on forecasting the weather?
"to recoup that they need to 'pander' to the most common platforms to see a reasonable ROI."
Someone doing profit/loss calculations at a government outfit?
Please clarify.
"They charge money for their an XML feed (which is basically giving you an FTP read access to a DAT file), that is why they won't have a XML/JSON api."
The format of the data is irrelevant, it is the data itself which matters and I am quite certain the Met Office has ways to monetize their data such that the basic info is free for powering weather widgets but the more detailed data is charged for.
The amount of data required to power a weather widget is likely to be fairly minimal - a location temperature, temperature range, wind, weather outlook, humidity for today and maybe up to a week ahead for up to 5 locations (for example). With updates occurring maybe every 1 hour. They could also have some reasonable terms of use such that a widget say "powered by the Met Office" and provide a link to their website.
More heavy duty apps which wished detailed information would have to supply a key with their request which is purchased and billed for in some way.
You, sir, are a moron.
Nobody's talking about forcing the Met Office to implement ZX Spectrum support, just wanting them to use an open standard (say, XML) instead of a closed one (Flash/AIR).
It surely wouldn't take a lot more effort to implement the client widgets in HTML/Javascript (as the widget engines in both Mac OS and Windows use anyway) and feed them from an XML file would it? That way, even if they choose not to develop a Linux widget themselves, the fact that the data to be consumed is provided in an open format means that anyone who wants to create their own widget is free to do so.
Met Office DataPoint
The Met Office already have a free XML and JSON API since November 2010, it's called DataPoint and will have more feeds on it within the next few months. For now see http://www.metoffice.gov.uk/public/ddc/. DataPoint is aimed at professional application developers, the scientific community and student or amateur developers, in fact anyone looking to re-use Met Office weather data within their own innovative applications.
'Development costs time and money, so to recoup that they need to "pander" to the most common platforms to see a reasonable ROI'.
How are they going to get a return on investment from something they give away for nothing?
The development costs come from taxpayers, so it's not unreasonable to expect that any taxpayer with internet access will be able to see the benefit.
So how much do YOU pay to access the Met Office Website?
What "return" does the Met Office expect to get from something it gives away for free, exactly?
"ROI" my arse.
As for your snide "Amiga, Atari ST or ZX Spectrum" reference, I bet there's a helluva lot more people running those systems under emulation than there are Mac users (currently only 6%, according to Hitslink), so why should the Met Office be "bending over" to support that platform, but not others, especially when an interoperable solution could easily be implemented using open standards, without resorting to proprietary junk like AIR®?
Not that the "ROI" from Mac users will be any more significant than that from other platforms anyway. After all, 6% of zero is still zero.
Looks like Adobe has just greased somebody's palm to boost adoption of their failing technology.
Screw the Met Office. Their "predictions" are useless anyway. I'd probably have better luck tossing a coin.
Macs have a damned sight more than 6% of the consumer PC market, which is the only market Apple have targeted until very recently. Unfortunately, nobody seems to be surveying that market alone as getting solid data for it is much harder than just letting a computer read HTTP headers from lying web browsers, servers, and all those PCs sitting in offices all over the world doing bugger all for 16 hours each day.
Almost every single consumer also has access to a work computer of some sort. The latter is the computer they're going to be sitting in front of for about 8 hours of every working day. Breaking out the figures that apply only to the consumer market is therefore tricky, although it'd be interesting to see what Apple's market share becomes if you only check data from weekends.
The Met Office isn't doing anything to support the Mac. If they're bending over backwards at all, it's to support Adobe's AIR. The complaint isn't the cheap flame bait of "oh look, the Met Office are supporting the Mac" but rather the one you almost hit — that they've decided to use a proprietary (and almost useless) standard for platform independence rather than an open standard.
But why can't they just release a API and let others sort out widgets and wongles and desktop pondles.
What about the blackberry tablet, seriously think of the children.
Mythtv's weather system uses XML files from the met office - under Linux.
Actually, it uses the BBC rss feeds. The BBC data is derived from met office data though, so it's really no different. There's no API for location searches so that part still breaks each time the BBC web monkeys decide to redesign it once a year.
So stuff them. Just use a widget from another provider. there's plenty of them.
The one on my google page isn't from the met office,and the only weather gadget I use on my linux system isn't either.
Why the fuck can't they just get the fucking weather report right and quit fucking around with other stuff?
They even have a iPad and Android app now, FFS!
Worst / Epic Fail is the latest version of the iPad app CRASHES, can lock up your iPad and is such an abortion of an app people are removing it en masse.
If you contact the Met's support email address, ipadapps@met.co.uk:
"We are sorry to learn you are experiencing difficulties with it and confirm we are aware that some users have experienced a service which has not met our usual high standards. The application is proving very popular, but there is a known "bug" which for some users is causing it to crash. Thank you for your patience whilst our developers are working hard in the background to resolve this issue. Full service should be restored without too much delay."
SCREW that, tell us accurately what the fucking weather will be!
Do the Met. Office, or original article, really deserve such a vitriolic comment ?
El Reg may enjoy ribbing them, but personally I've found their predictions for this last year to be spot on.
That is, after all, what they're there for. Apps and widgets are just gravy.
He does have a point though - the only thing they seem to get right these days in the UK is the damned temperature, and that's not always right either.
The one they simply can't seem to get right is precipitation - unless you're on the west coast of scotland, where you don't need to have a meto. degree to forecast rain. Every day.
Though for balance, he does sound like a man who gave up smoking 3-4 days ago...
It does seem as if the Met Office manage to screw up just about everything they touch. Their iPhone app contains horrible bugs that make me wonder what their developers are smoking. Every time they update their web page, the usability deteriorates still further. I just can't understand how they're managing to get these things so wrong, unless they're outsourcing to the cheapest and most incompetent bidder.
The only thing guaranteed is rain, but like everything else 80% of the time it just pretends to do that too.
to look at atlantic isobars - as a 'weather' site is been so dumbed down its useless - almost as bad as Metchecks semi-permanent 'Subscript out of range:'
Sorry Met Office but www.yr.no seems to be rather more accurate than your forecasts.
I agree and their xml based api is available and documented.
hahahahahahaha
yr.no is the least accurate weather site out there. Really. They're utterly useless. About four years ago I thought they were really accurate - that might have just been my perception back then - but not now.
The really funny part, however, is that they get their forecasts straight from the UK Met Office. They had a major failing about nine months ago where their >3 day forecasts were totally out of whack. They were blaming the fact that the UK Met Office models couldn't cope with the mountainous landscape that features so prominently here in Norway.
One of my android widgets uses yr.no... Yesterday, it claimed it was thundery weather all day (all it did was rain). Today, it's snowing and all it's saying is rain.
Useless. Can't even get the CURRENT weather right, let alone predict it.
I've never really understood why it is considered necessary to know what the weather will be from minute to minute.
I can understand that an accurate forecast is very important for certain occupations and industries, but does the average office bound worker really need to know there is a 10% chance of light rain in the next hour?
For most purposes, a quick look out of a door or window will suffice.
My own work involves my being out of doors regularly, and I use the Met Office's site once a day to look at the surface pressure charts and forecasts. This is mainly to make sure I have appropriate protective clothing in the van. Popping home to change, is not an option.
As someone once said, "it's not bad weather - you're just wearing the wrong clothes."
Do I need to explain the icon?
"I've never really understood why it is considered necessary to know what the weather will be from minute to minute."
"I can understand that an accurate forecast is very important for certain occupations and industries"
So, having said you don't understand it, one line later you do. Excellent.
...percentages are wasted on a lot of people. If you tell them there's a 80% chance of a harsh winter and then it's mild they'll say you "got it wrong". Of course, the same morons keep on buying lottery tickets...
Sorry to have confused you. Perhaps if you read my second paragraph in it's entirety, all will become clear.
What I was trying, and probably failing, to say was that most commercial users will be making decisions on forecasts for periods considerably longer, or further ahead, than the next few minutes.
Hence my puzzlement at the need to know the weather "minute by minute".
"What I was trying, and probably failing, to say was that most commercial users will be making decisions on forecasts for periods considerably longer, or further ahead, than the next few minutes."
Ahh, fair enough. Well, as both a keen ultra-distance race-walker and former Mountain Rescue volunteer I can assure you that the hour-by-hour forecasts are hugely useful. Gave up on the met forecasts 4 or 5 years ago, mind. Nowadays I read the pressure charts & radar to get an idea of where things are and where they are going followed by Accuweather for a pretty good stab at what will happen. Their granularity is also much smaller than the Met (i.e. they claim to offer forecasts for smaller areas) and I usually find them to be spot on.
Not holding out much hope for accuracy this week, though. The pressure charts show the warm front edging back and forth over where I live - which basically means prepare for all types of weather every ten minutes or so.
adobe have been a bit vicious by actually removing the adobeair package from their yum repository. Frankly, I don't see what harm there was by just leaving it there ad infinitum. The binary install provided on their archive page is a bit of a nightmare to install on modern Linux OS's.
If find it works very well, and I'm fairly sure Auntie gets her info from the Met Office. Problem solved.
Here's my Norwegian forecast:
Cold. Wet.
Because that's where the weather is!
When a Windows app causes a crash it's the OS' fault. When it's a nix-based OS, it's the app's fault.
It really matters very little just what the interface is, even that "simple json or xml" mentioned above is easily overkill. There already exist weather report and forecast exchange formats that predate the webmonkeys' latest and greatest formats*, and fetching that might already be enough and lower traffic than either. What matters is the thinking**, the understanding both of the demand and of the expected ways to solve it. The met office fails at least on the latter, missing a chance not only to deal with adobe's high-handed refusal to support any open platform for their air product, but also to garner reputation and support, facilitating innovation, and so on, and so forth.
They could've done this right by starting at the API level, whatever its form, publishing one or more usable interfaces with documentation, and writing their own widgets against that. They could even publish the source to those just to show how it's done.
But to do that they'd have to understand that putting the weather on the intarwebz entails more than providing code that only runs on the platforms some provider of plugins deigns to support. Letting yourself be led down the garden path this easily is not a good show. It merely sets you up for more of the same kind of failure a little down the road.
* Note how xml is so great that people came up with json in response, and neither has managed a monopoly. Neither have the lesser-known alternatives. How convenient.
** A nicer example is how the dutch met (knmi) publishes small aviation weather reports in a boring old text format stuck between a pair of 'pre' tags, surrounded by html fluff. I have a script that fetches the page and displays only the part between that tag pair. I'd waste a little less bandwidth if they'd publish just the useful text part.
You are not going to emit whines about writing two additional lines of Perl, are you?
The script is a shell script, mostly scaffolding to provide easy extensibility and usage output. The pre filter is two lines of sed. Adding in perl would mean a rather larger footprint, and json/xml parsers add further dependencies of actually quite a bit more than two lines.
We blithely assume that coder time is more valuable than cpu time so tend to throw more and more code at the problem to make the problem vanish (either through solving it or trough obscuring it), then throw as much hardware at the resulting mess to make it look like it runs acceptably fast. That is quite a lot of wastage, oftentimes easily remedied with a bit of stepping back, looking at the thing in its entirety, and thinking for a bit.
Not sure if that's a whine, and I wasn't about to bring it up, but then you did. This script I wrote for my own enjoyment so it's not critical; if it was more trouble I probably wouldn't have written it at all. But time and again, and both within computing and without, you see that being careful at the start of the pipeline, even with seemingly minor details, can easily yield substantial benefits at the end. Computing is full of such complex systems that we tend to automatically compartimentalise in modules and layers and whatnots, then possibly optimise within the divisions we imposed, but easily forget about the complete picture.
The result is a bit myopic and one easily overlooks even simpler and more efficient solutions that might not merely be possible but already exist. That's what I was on about, regarding the blithe "must have json and/or xml!" comments a bit above. Your "two lines of perl" is a bit of a red herring, perhaps because you yourself missed the point. This particular field already has acceptable transfer formats, why not use them?
In a way, that blind preference for what aren't even very good general formats is the same sort of fault we're accusing the met office of, though perhaps less cripling (for us freetards). And then there's the habitually understated cost of using them.
Excusing the fact that your anonymous here... (clearly a paranoid programmer)
I like the cut of your jib.
I've always found XML and JSON quite excessively verbose, whilst understanding that the underlying principles behind them are sound - standards and data freedom.
It's all about what is best for the job at hand - and if that's just a simple blob of text that can be grabbed by wget and formatted for presentation = job done.
All that matters is that the data source retains it's structural integrity.
It so ridiculously easy to take that exceptionally basic data structure and wrap anything you want around it - pretty icons? well, you've got the cloud and temperature data, what more do you need to extrapolate that data into an image?
10 lines of Javascript could take that basic text feed and output a relevant weather icon, or 10 lines of php, or 10 lines of Perl, or 200 lines of asp... (low blow)
Obviously at the "Foreskin^H^H^H^Hfront of Modern Technology"
