I am not suggesting anybody does this because we already have plenty on the market, but what are the biggest challenges?
Years ago, I understand it was quite easy to get one started, by getting the free DAFIF database from the USA. DAFIF was populated by the various national CAAs around the world and it was wholly machine-readable.
It was killed to public access around 2006, when various CAAs (starting with Australia I believe but also including the UK CAA) threatened to stop supplying the USA with the data unless the USA stopped publishing it - on copyright grounds. Jepp got sued around that era also, and reached a confidential out of court settlement.
When DAFIF went dead, most of the US software vendors whose products included non-US data withdrew support for non-US regions.
Nowadays, I think the only way is to systematically browse the national AIPs, but they are not machine readable. The best one can do is swipe the text in the PDF and copy/paste, and perhaps script-automate parts of that, which will work for a bit... Must have been fun for say Greece whose AIP was not available electronically till very recently, so one would have to scan and OCR the book.
Recently I heard that the UK AIP delivers the Class A airspace shapes not as the shapes seen on the chart but (partially) as airway centrelines and widths, and the software has to merge these and correctly work out the intersections in 3D.
The other part of the job is airport data. Some of this is in the national AIPs but many of those are incomplete on stuff like contact details. Years ago, when I used to fly VFR with the Jepp Bottlang guides, I used to find loads of completely duff phone and fax numbers. Navbox was a lot better, but the question remains: since there is no reliable source for this data, and since no software vendor is going to manually verify every contact phone/fax/email, how do you deal with issues like user reports of wrong numbers, where the right number is provided but the next AIP edition is bound to replace the right number with the previous (wrong) one again...
I suspect that if one started to dig in it one would find a way. However, making anything for aviation is talking to 5% of the total market.
Back in 2003 I made a mobile weight and balance application (Flight WTK) running exclusively on mobile phones. Although it sold fairly good (UK, France, Spain, USA) it never sold in huge quantitiies. After two years or so a US company bough it from me. Don't know if they ever did anything with it but I know Aircraft Spruce sold copies of it.
However (finally, the point ) in retrospect, I wishg I rather had made a program for the whole market, and not only the 5% flying small airplanes.
Parsing AIPs is what vendors (including Jeppesen) do and you can automate this to some degree but have to be prepared to constantly adjust your algorithms and do manual work.
Until some years ago, the German DFS (which is a government company but required to partially fund itself) intentionally delivered the AIP to Eurocontrol in bad quality fax transmissions. They still withhold the VFR portion to protect their business.
The flight sim people have fairly good databases which are open but of course without any guarantees. Unless it was terribly important to be correct, I would probably use the sim databases.
As to doing business in GA -- not something I would consider. This market meets all criteria for a really bad market: it's small, it's declining, heavily regulated, conservative and there are lot of hobbyists around with too much money that think aviation is cool and start a "business". I have seen so many "companies" that only serve the purpose of turning the owner's fun flying into a tax deductible expense. In a declining market, you have to massively take market share to grow. In a growing market you don't have to do much to grow.
Regarding phone/fax: that was a problem before Internet became widely adopted. Nowadays hardly an issue.
A very good analysis, Achim - I did consider having a go in this market but soon enough I came upon every truth you put forward. It remains a nice hobby project, but I've given up on commercial ambitions in this respect. As you said, too many anglers in a pond growing only smaller day by day.
[[afterthought: but is the opening question well worded? didn't it rather mean "how would you create and reliably maintain a database of european airnav data, like aerodromes, airspace, navaids, &c" ? ]]
As I said, I don't think there is any point in trying it. The market is very fragmented, presumably because few pilots fly internationally. A colleague one showed me a superb app for Germany, but it was all in German and for me totally unusable.
My curiosity was how does one maintain the database in the long term, when there are no obvious data sources.
One can get quite a lot of stuff from Eurocontrol. I haven't been to the EAD site for a while but one could print off various reports e.g. a list of all DMEs and their parameters including lat/long.
Eurocontrol will also sell you a B2B access package, for a few k a year I believe, which gives you access to the EAD database via a VPN, and presumably there is an implicit permission to use the data commercially. Those reports can be easily machine read.
So some of the info can be obtained more or less automatically.
But airspace info can't. Eurocontrol have a curious free product called Skyview which comes with an IFR database for Europe, but the database (which is in a standard database format) is always out of date, just a little bit.
How would one handle user reports/corrections if they conflict with the data which you are periodically grabbing from "official" sources? I don't think any existing vendor deals with that.
I still cannot see any place from which one could grab VFR data.
I don't suppose there will ever be any, since all (well, most, e.g. Greece doesn't) European CAAs print VFR charts as a profit centre.
Skyvector do AIP parsing, just like the others. A very good presentation.
The UKs class A airway system seems to be entirely missing from there.
Here in Germany, the basis of everything (so to say) is an official gazette called "Nachrichten für Luftfahrer" (which translates to "notices to airmen" but is different from NOTAM!). There is an online subscription available with updates every two weeks. I am pretty much certain, that a similar publication exists in every country. Short term changes are published by Notam (including changes of phone numbers and even airfield websites). So I think one needs to constantly evaluate both sources of information to keep an up-to-date database.
Not worth the effort, too much competition from long established companies many of which struggle to make a profit.
Any of you looked at RocketRoute lately?