Re: Is this a good analogy?
". . . an impossible to enter house is essentially impossible to build . . ."
And this is why justifying such measures by way of analogy is misleading. There's the same argument about metadata being "the name and address on the envelope, not the contents of the letter".
Analogies are fine for understanding non-critical concepts - they give you a way to explain/understand something that is difficult to grasp directly. I do this daily in my discussions of IT with non-technical people and that's works well.
But, to explain important issues that contain complex nuances with wide-ranging consequences that will affect vast numbers of people and companies in diverse fields in myriad technical, logistical, developmental, financial and legal ways, well, 'think of encryption like a house' just doesn't cut it.
If it must be compared to some more familiar, physical object, compare it instead to a combination safe. This has and instant benefit over a comparison with a house door in that a (proper) safe is designed explicitly to prevent it being forced open by unskilled people, whereas a standard house door is really only designed to keep out casual intrusion and cold breezes.
Now, there are many such safes and some are stronger than others with varying levels of complexity and combinations. Safes are, as can be implied, a safe place to store things - not just money but information, passports, personal memories like photos - whatever it is you want to keep, well, safe. And not just safe from thieves but private as well. Perhaps you have correspondences and keepsakes from an ex-lover that are still dear to you but that you don't want your current partner to see. Or cigarettes or a bottle of nice whisky that you don't want someone else to pilfer or find out you've been smoking/drinking.
And, of course, companies have safes too - usually for money (or equivalent) but also for things like backup drives that they don't want lying around or for copies of the company ledger.
The encryption algorithm/method is, in this instance, the workings of the safe locking mechanism and an encryption key is the code you put in to unlock it. With this analogy, what the government is asking for is for EVERY safe to be able to be opened by law enforcement, without having to actually ask the owner for the combination.
Think what that means for a moment . . .
For a start, we have to come up with a mechanism for this unlocking to occur. There are three main ones that are available.
First, one can require that everyone who owns a safe must provide the government with the combination. This pretty much requires that you couldn't just go out and buy a safe - you would have to apply for a license to own a safe and register it with the government. You would then be required to update the government every time you changed the combination on the safe. Otherwise, a criminal could simply go and buy a safe and just not tell the government.
Second, one can require that all safes have a second, fixed, code that is specified by the government or supplied to them by the manufacturer. You could have one code per maker or per model*.
Last, we can require that all safes have a 'backdoor' - a mechanism of opening them without knowing the combination.
Some may realise that this last options is what most actual safes do, in fact have and the knowledge and ability to 'break in' to them is closely guarded and only provided to approved, accredited locksmiths. There are caveats, however, such as the technique being specific to each model and some of those are destructive, requiring drilling in precise locations, usually using templates. There is also the fact that no safe is actually required to have such a procedure and all such procedures take experts with specialised knowledge and non-trivial amount of time (and hour or more, usually) and generally it's pretty obvious what's going on.
But even then, with all that comparison, there are still CRITICAL points of difference, such as safes requiring individual attention of an on-site person. I.e. it is not feasible to 'break into' many safes simultaneously or to break into one remotely and usually not without someone knowing you're doing it or have done it.
And this is where any analogy falls down, because none of them come close to either the breadth of access that 'crackable' encryption would allow or the ease of an 'authorised' person doing so or the scope of how many people could be affected simultaneously or the ability to do all that without anyone knowing.
Feeding that access back into the safe analogy, the access they are trying to achieve is not just to be able to break into any safe they want but to be able to remotely, secretly and nearly instantly determine the contents of every safe, owned by every person, store, company, pub, rotary club, church, bank, school, oil corporation. Every political action group, every civil rights organisation, every media outlet. They want to be able to find out which pornos a 17 year old has stashed under the mattress and what's written in your daughter's diary.
But more than that. They want to be able to record it all - not just what is in there but when you put it in and when you take it out and if you transfer it to another person's safe, whose and when they take it out and who they transfer it to. They want the ability to invisibly copy the the contents of your safe - unknown to you - and keep that information forever.
But it doesn't stop even there because this ability, to break encryption seamlessly and without anyone knowing it's been done, also allows someone to alter the contents of your safe when you're not looking - to remove a photo you've stored or to corrupt a document so you can't read it anymore. Or to add stuff in.
And that's worse and the analogy can't keep up, even stretched as it is, because the abiity to decrypt your information allows for 'man in the middle' attacks which, in concert with the existing ability to interecept communications can alter your data in transit if so desired. A file downloaded from from a website could be replaced with what would appear to be the same file but was infact altered to infect your computer or device with malware - a key logger, for example.
The implications of this are just mind-boggling and no analogy, no matter how relevant it might seem, can capture the full scope of what is being proposed. Any attempt to explain it in such a way risks misleading or, is designed to do so.
No, metadata is not just like the address on the envelope and the ability the government is arguing for is not like being able to enter a house. (Regardless of warrant.)
* - To keep the analogy in line, we can imagine that individual safes can't be identified - for example by a serial number - and so a per-unit code hard-wired code is impossible.
Biting the hand that feeds IT
