This is just publicity for http://www.originstamp.org/ ; skimming OP, the only thing they seem to do differently from the obvious 'hash, turn hash into Bitcoin address, send satoshi' is they batch up a day of submissions into a single hash/address for efficiency:
> To reduce operating costs, the server collects submitted hashes, concatenates the hashes, and
generates a single aggregated hash. By performing additional hashing and encoding operations, the
4
aggregated hash is converted into a Bitcoin address . To this new address, the smallest transactionable
amount of Bitcoins (0.00000001 BTC) is transferred. Each transaction requires a fee of currently approx.
0.0001 BTC (3 US cents). By including only the aggregated hash in one Bitcoin transaction per 24 hours,
the total transaction costs are less than 10 USD per year. Users are given the option to include their hash
in a transaction that is performed immediately for a fee of 1 USD. To provide non-paying users with
immediate evidence, their hash is published to Twitter right after submission.
The biggest difference is they published a paper on it, which I guess is good marketing ( I hadn't heard of proofofexistence.com).
More substantively, they pool transactions and eat the cost, so it's free for the user. proofofexistence.com charges the user 5 mBTC, and you need to know how to send bitcoins to an address. Also they use Twitter as a secondary hash store and provide a somewhat better interface (you can directly upload hash in case you don't trust their javascript hasher or directly type text into a box, and they provide a git plugin as well).
http://virtual-notary.org/ has been in this business since 2013. It allows more than simple document uploads and will notarize weather, exchange rates, Twitter feeds, etc.
The reason for that limit is that the bigger the completed block, the longer it takes to propagate. If it's too big, it takes too long and you get network forks.
Right now they're working on a solution. Since most of the miners already have most of the transactions, you can propagate a small fixed-size block that lets them reconstruct the full block from the transactions they already have.
This fixes the delay and the forking problem, so you can put many more transactions in the block. Your limit is just the bandwidth to transmit the transactions. Typical home broadband can support tens of thousands of transactions per second.
The storage of old transactions is another issue, which they plan to address by pruning old empty addresses.
