Of companies mentioned in the article, the asymmetric contributions to OSM that bother me the most are AllTrails and Strava - 2 companies that heavily rely on trail data sourced from OSM.
AllTrails does close to zilch to help put trails into the maps, even though the majority of trails are user-generated content.
Strava has let "Slide", their one project that could help put trails on the map, die an ungracious death.
If you build a whole business model on top of free data, it may be worth considering improving that data.
Strava worked with Mapbox when they were switching over. I don't know what exactly the deal was, but Mapbox employees added a bunch of missing stuff to OSM based on Strava data analysis.
but I do find this whole discussion strange, because it's rather obviously the same problem that open-source _anything_ has: you're giving something away for free, and it's very hard if not impossible to control how that data gets used or whether the primary benefactors contribute back anything to the source.
Ironically, the only known model of forcing people/corporations to contribute something back based on their usage of something is called "market-based pricing".
Unless someone is proposing that maps are somehow ethically or systematically different, than, say, Linux, this conversation feels rather unspecific and pointless.
There's a similar problem with websites which map rock climbing routes. Some make their data open to some extent, but not in a way that can be meaningfully contributed back to OSM. Even though they all use OSM data (usually via Mapbox), to generate their maps, and allow users to draw and annotate layers on top of it.
I'm very interested in working towards an open tooling and open data ecosystem for rock climbing information, but I don't know where to get started on finding others to build it with.
Does strava actually use the trail data from OSM? as far as i can tell, the routes on strava are entirely contributed by strava users and don't come from OSM in any way.
The two maps follow the same paths obviously, but at least in my region the routes and segments on strava don't ever seem to start or stop at the trail intersections in the OSM maps and often cross unmapped and unofficial connectors that don't appear on OSM. The extent of their reliance on OSM trail data seems to be that they use MapBox tile images and those tiles sometimes have OSM trail names marked on them.
Strava's map are based on OpenStreetMap, there is attribution in the bottom right corner of every map (not sure where in the mobile app, I'm not an active user).
The parent comment probably talks about giving Strava user uploaded data/tracks back to the OSM community so they improve the map (or add new paths themselves). In a way they already do https://wiki.openstreetmap.org/wiki/Strava#Data_Permission_-... Allowing tracing is more than other companies offer and gladly accepted. It regularly leads to new unmapped paths discovered, sometimes on private ground (military areas) that are otherwise inaccessable to an OSM mapper.
yeah, that's what i was saying about the tiles - their tiles come from mapbox, which means they include OSM data and therefore must include the attribution. but the claim was that they're heavily reliant on OSM's trail data, and I just haven't seen any evidence of that.
Actually, yes. We offer it as a pilot for folks who really want it and are willing to put up with some bugs :) Just text your bookkeeper "HN 2018 tax prep" and we'll set it up for you. It takes 1-2 weeks.
Liked this post. Would love more detail on GPU query offloading and on how different the engine looks compared to one that would run queries on the CPU, as that seems to be the primary innovation
Shoutout to PiHole - https://pi-hole.net/ which falls under the category of DNS blockers. One tiny Raspberry Pi can provide adblocking to all the devices in a network and radically speeds up page loading.
True, you can't block trackers on domains that aren't blacklisted, but on the other hand you can block trackers for all devices on the network.
I had Ghostery installed for years (gets one of the highest blocking marks) but I didn't like that it slowed down Chrome, hogged a lot of memory, and has the capability to send traffic back to a company.
A lot of startups get sued by Non Practicing Entities (trolls) that have no operating business to speak of. That provides the slimmest counterattack surface - so not sure how useful it is to have a portfolio to hit back.
I think a better defense mechanism would be a legal defense fund that would force the trolls to make their case at trial, to go through discovery, to bring expert witnesses.
Because the law firms that represent them work on contingency, this would effectively cut into the potential licensing fees, and make trolling less profitable (and less likely).
Also, a lot of trolls extract patent licensing fees out of startups for patents that should have never been granted, and that deserve to be invalidated (e.g with prior art), but the process of getting a patent thrown out is expensive, so having a fund would greatly help.
The other strategy that could work is to get all the startups that get sued by a troll for a specific infringement and make a sort of "reverse class action" - making it possible for the startups to re-use the same lines of argument, evidence, etc in their cases.
Others have mentioned this, but this PatentSheield scheme is not useful against non-practicing entities (NPEs or "trolls). It is intended to be used by startups who are at risk of sued by big incumbants. For example, if an IoT startup participates in this program and is sued by an established IoT company, the startup can take title to the portfolio and counter-sue the incumbant for infringement of the portfolio patents. At that point, the companies agree to a cross-licensing deal and both parties live to fight another day.
This kind of arrangement is not useful for countering NPEs because NPEs are non-practicing by definition. Since they are non-practicing, you can't countersue for infringement.
Your note about cooperation between potential defendants is spot-on. The difficulty is in getting them to cooperate. Accused infringers are likely to be competitors who may not be inclined to cooperate. Further, it is to the advantage of each to simply license the patent and wait for another accused infringer to do the work to invalidate the patent, at which point they can all stop paying license fees. (A classic "free rider" problem.)
By the way, a company can buy patent infringement insurance that will pay for litigation or settlement costs, but it isn't clear that it is worth the premium.
Let's say I pay into this fund, then I get sued for patent infringement. How does the fund decide whether this lawsuit is trolling or legitimate? Or would the fund pay for all patent lawsuits?
There are several ways of reaching a distributed consensus about that, like proof of stake.
It is also conceivable that it would be in the fund's interest to defend all lawsuits. That would provide the largest deterrent, since the lion's share of today's suits are without merit, and the cost of defense would be considered the operating cost of the fund.
What the trolls exploit is the assymetry between a defendant without resources and their own resources which are solely dedicated to this.
Without all the licensing fees that startups pay to settle the alleged infringement the business model for trolls would stop working.
That would provide the largest deterrent, since the lion's share of today's suits are without merit
Doesn't this depend on what "without merit" means? It seems unlikely they are "without merit" in the legal sense, since then they would be thrown out at the first hearing.
If you mean they are from patent trolls aka non-practicing entities, then it seems wrong too. I see coverage of tens of cases by NPEs per year, but the entire patent law-suite field must be much bigger than that.
There are probably better sources than this PWC report[1], but it identifies 477 out of 2281 (ie, 21%) cases since 1996 in the districts where NPEs file cases involving NPEs. Their definition of NPE includes universities though, which I don't think is what people generally think of when they think of "patent trolls".
If "without merit" means that they almost always lose, then this doesn't seem right either. The success rate for patent cases varies, but seems around 25-35%. That's far from "lion's share without merit"
Ars Technica reported last year that in 2015 over two thirds of patent lawsuits were filed by trolls and that of all patent lawsuits, 44% were filed in the Eastern District of Texas (where few others than trolls would be filing suits). The report says that although the numbers are up, the trolls' business might be down.
The total number of cases seems to line up between the two reports (5700 in 2015, not including PTAB).
But only 2281 cases involving NPEs were decided between 1996 and 2015[1]. If 66% are filed by NPEs, then this should be much more. Unless the vast majority are dropped I guess?
Yes. Cases issued will not correspond to cases decided, since many cases settle. And there is a long time-lag from starting a case to ending it with a judgment.
> It seems unlikely they are "without merit" in the legal sense, since then they would be thrown out at the first hearing.
In all patent troll cases I've seen, a patent exists, and the defendant is actually infringing the patent. So there is some merit to the lawsuit.
The problem is that the "inventions" in the patents:
* are not novel, or
* are not non-obvious, or
* have not been described in sufficient detail to replicate without further research, or
* a combination of all of the above for various claims in the patent.
so the patent should never have been issued in the first place. But invalidating the patent in court is a tedious and expensive process. You need many expert witnesses debating what is novel, what is non-obvious, and what is sufficiently detailed, it's not something a judge can decide in 10 minutes.
Couldn't the fund just use arms-length consulting patent attorneys to answer the question: does defending this case have a reasonable prospect of success?
It could. It could also force the startup to hand over control of the defense (which the startup would most likely be ill-suited to control) and therefore control the defense strategy.
All that is required is a big enough defense fund that acts as a true deterrent.
Mind you, if the startup loses it doesn't mean that it would not be responsible for the payout - but at least the startup wouldn't fold simply because the costs of defending itself are out of reach.
I see this type of refining activity on our payment gateway logs. I can only imagine what it looks like at a major online retailer or SAAS service. It just goes to show how fundamentally broken is payment card processing at the moment, and how ripe it is for disruption.
My guess is you're talking about the logs from your business site, Ringio.com. The "refining" problem (testing stolen credit cards) is much worse for sites that generally receive small, one-off payments. One of the hardest hit groups is non-profit organizations who accept/solicit donations. They regularly get clobbered by charge-backs and other issues.
The charges against Ellsberg were dropped when it became public that the government had spied on Ellsberg by wiretapping Ellsberg's psychiatric visits.
So maybe all Snowden has to do is to show the government spying on him. Which would be kind of meta, as it would put the government on trial for PRISM itself, while probably acquitting Snowden
Just to point out that when you send a PDF document signed with Preview, I can open that document, extract your signature, and paste it in any other PDF that I want to. For me, this makes it too dangerous to use.
To me, the article is about 1 or 2 years too early. It provides great detail about all the pros and cons of judging the choice in advance, but no evidence that they made the right choice. Unless Spolsky is ready to start judging the performance of a sales team by some metric other than revenue.
I hate scanning. What a waste of time. I simply subscribed to http://shoeboxed.com, which syncs my scans to evernote and the other programs I use to manage the documents.
So pretty much I gather all the paper, put it in an envelope, and mail it once a month.
AllTrails does close to zilch to help put trails into the maps, even though the majority of trails are user-generated content.
Strava has let "Slide", their one project that could help put trails on the map, die an ungracious death.
If you build a whole business model on top of free data, it may be worth considering improving that data.