Wednesday, December 10, 2014

Evernote, Forevernote, Nevernote?

I write a lot of things down, most of them on a computer. What I’m writing on a computer usually falls into one of several fairly typical categories:

  • Brainstorms
  • Plans
  • Lists
  • E-mails
  • Software

I wish I could add essays and fiction but I’m not there just yet. And I do keep something vaguely like a journal, and something like a studio notebook, but those are done with pens on paper and thus, as we say, out of scope. I might yet get around to regularly blogging or otherwise writing for the web, but I’ll burn that bridge when I come to it.

So, back to my categories. Let’s start from the bottom. Software, for me, lives in GitHub as long as it’s my own software. If it’s work for hire then I’m not at liberty to say where it lives. And most software is written in an editor of choice. When the choice is mine it’s usually TextMate 2, otherwise it’s usually XCode; and of course Vim is the fallback. Thus you could say that the software-writing problem is solved, for better or for worse.

I would say the same about e-mail, though here most technically minded people would aggree that “for worse” applies. I write my e-mails in Apple Mail and that’s unlikely to change. If I were a Windows user I’d probably use Outlook, and if I were on neither I’d probably use a mix of Thunderbird and something like Pine. I’ve used all of these as my primary e-mail client at some point; I’m not a Gmail type, nor do I “glass”, nor do I drive a Segway. Call me old-fashioned, or say I have a modicum of taste, or ride into the sunset on my color TV screen. As you like.

At any rate I, like most people who write software and e-mails, am sufficiently invested in my current workflows that it would take something fairly revolutionary to make me switch.

The next thing up my list is a different story: Lists! How do I list thee, let me list the ways! As banal as it might seem at first glance, list-writing and list-managing and list-using are all basically unsolved problems that are perfectly suited to software. And because nobody has really solved these problems together in a universally good way, and because they look on the surface like simple problems, there are lots and lots of computer programs that try to help you with your lists.

Most of them, as far as I can tell, are concerned with the subgenre of To-Do Lists. Among those, one of the most established and probably the most comprehensive is OmniFocus, and I have used it for years. I bought their Mac software and their iPhone software and their iPad software, and thanks to Apple’s dislike of “upgrade pricing” I may well buy them all again.

OmniFocus is GTD-oriented but it works for generalists just as well. I think it’s fair to say that OmniFocus was built for, and is championed by, people who have resigned themselves to always having large and complex To-Do lists, and in general to being “busy people” in a way that involves, or at least should ideally involve, lists. I may be such a person, and that may be a good thing, or maybe not and/or not. That’s a topic for a more philosophical blog post in The Future.

Over time I developed a serious problem with OmniFocus: it was such a good program for “capturing” (writing down) list items that I started keeping all my lists in it, not just my lists of actionable To-Do items. That was very much not the intended use of the program, and the user interface frequently reminded me of this fact.

Let’s move further up my category list: more trouble arrives with Lists. My List-Landfill slowly morphed into a Plan-Landfill: I was using OmniFocus as a project outliner, again because it was so easy to put things into it. But a Plan-Landfill – Plandfill? – is a dangerous thing. After all, these are plans! Things to be realized! Not things to be forgotten!

So after a year or so of this I decided to move the plan-writing to OmniOutliner, which seemed like just the ticket. And I suppose it might be, but I couldn’t warm to its look and feel, and its file format and export options were adding complexity where I already had enough.

Meanwhile I was still using my favorite text editor, TextMate 2, to do most of my Brainstorm writing –- the top of my writing pyramid, or column at least. I would write things in Markdown, and when then got too long or needed images or diagrams I would break them up into messy directories of Stuff.

That is, when I wasn’t further abusing the OmniGroup by mixing up my brainstorms and my plans and my lists in OmniOutliner and OmniFocus. Oh recidivist I!

Eventually it dawned on me, like sign language on a chimp, that I was overcomplicating things and that none of this would ever “scale” or for that matter remain useful five or ten years down the road.

As it happened OmniFocus had just been upgraded, and it would cost some money to re-commit; and I had been growing weary of their sync service that almost never worked with my iPhone (I had to re-download my whole database several times a week). So I decided to try something new.

Enter Evernote.

But wait: first, enter Wunderlist. It occurred to me that the best way to stop overusing OmniFocus was to use something else for my To-Do lists, and there was a new startup on the block with a well-reviewed and simple app. As a bonus they are based in Berlin, as am I (more or less). Wunderlist’s business model seems to be team collaboration and subscription revenue, and they’re doing a good job as far as I can tell. As usual, I digress: Wunderlist will be the subject of Yet Another Future Blog Post.

Enter again: Evernote.

The promise of Evernote is to be, like Yojimbo before it, a catch-all information organizer. Only modern. On all the smart devices, in the “Cloud,” and so on.

This works pretty well in practice. You have “notes” (documents) which are organized in “notebooks” (which can be nested), and there’s some basic tagging support and some interesting takes on organizing it all in the UI. Notes are more or less rich-text, but can have quite a bit of styling in them, and you can add images and other media. You can “clip” (capture and save) web pages. You can graphically annotate your graphics, and your notes as well. You can share notes and collaborate and this and that and the other popular thing.

Overall, I like it, especially for brainstorms.

Evernote works on the subscription model much like Wunderlist, and is also pushing its collaboration and “business” features a lot lately. It’s free to use and quite useful at the free tier; the paid version is something like $5/month, which for us grownups is more than affordable. In fact as long as you live in the developed world it’s affordable: it costs less per month than a single Starbucks coffee-esque beverage, and it provides real utility.

As with any app I disagree with some of the UI conventions, and as with any app there are some annoying bugs, but overall it strikes me as a great product, and after a month or two of use I could see it potentially solving my brainstorming, planning, and list-writing problems. It even supports To-Do lists, but based on my experience with OmniFocus I’m keen to keep that function separate from the rest.

Evernote has been very successful and as far as I can tell the success is well deserved. It even has a public API that a hacker like me could use to integrate it into other, weirder parts of my information superduperset. Part of me wants to go all-in and make Evernote my primary note-taking, information-ingesting, brainstorming-and-more gateway to the magical Cloud in the electric-blue sky.



But what about the lock-in? And what about security?

To be fair, I think Evernote goes above and beyond the industry standard for both data-freedom and security. There is a well-documented, machine-readable document format and you can export all your data whenever you like. You can even do it through the API if you want.

The company clearly takes security seriously and does a lot, probably all they reasonably can, to keep your notes secure and private. Or at least all they reasonably can while maintaining a useful service for normal humans.

As I see it, the lock-in problem for any system comes down to two questions:

  1. If I want to move to another system, how hard will it be?
  2. How confident am I that the service (and my data) will be around (and useful) for a long time?

Moving to another system will be easy as long as the other system imports Evernote files. Given Evernote’s popularity I think it’s very likely that serious competitors, once large enough, would import your data. Or at least import my data: in my case it’s mostly text, so I’m not worried about importing a terabyte of painstakingly catalogued home video.

Moving to no system, or to a system with no Evernote import, is another matter. At least it could be done, but I’d probably spend a week writing that converter. And the thing you’d end up with is, at best, a big pile of HTML documents.

For most people that’s utterly irrelevant. But for us hackers it’s a thing to consider: the portability of a document decreases as its structure gains in complexity. Evernote’s HTML variant is pretty manageable, especially by software, but it’s still a far cry from the plain-text goodness of Markdown.

I don’t think anybody’s doing this better than Evernote, but it does give me pause that there isn’t a simpler underlying format in case I change my mind.

And what of the product’s longevity? This matters not just in terms of access to your data but also, principally, for your continuity of habit. If you learn all the tips and tricks of Evernote, build yourself a workflow around it, and use it day in and day out, your commitment has great value. The value to the company is obvious as long as they keep enough customers paying; but the value to you is much greater. The cost of learning another system would be very high.

Plus, as unpleasant as it is to think of such things, longevity isn’t just a matter of existence. You need the product to be cared for, to be developed futher with the kind of passion that’s brought it this far; and above all you need them to keep caring about data-freedom and security.

In favor of Evernote’s longevity we find, mostly, their great popularity. I think their business model also makes a lot of sense: get people involved and committed on the free tier, sell them up to the premium tier, and make most of your money from a business tier that becomes more and more attractive to businesses as more of their employees already know the software.

But I don’t think there’s any proof yet that I’m right. They may well have a very lopsided user base, with the free tier consuming most of their resources and the paid tiers not contributing enough money.

They’ve apparently raised over $290 Million in venture funding over many rounds. That’s enough that there will definitely not be an OmniGroup-like small-shop happy end to all this. Either Evernote goes public, and is subject to market pressures that may force it to change its behavior; or it is acquired by a large player like Google, and is “integrated” into something like GooglePlus.

Or, of course, they go out of business. Given their open data format they are more vulnerable than not should a game-changing revolutionary note-management thingy evolve and come shuffling up out of the mud. I don’t think that’s likely, but then I was one of the many geniuses who thought up Twitter in 2005 and then dismissed the idea as culturally pointless and commercially hopeless. (We may all be proved right in the end…)

All of these risks exist with any product you might choose, certainly with any product that’s this popular. I have no reason to believe the odds are any worse with Evernote than with any of its competitors.

But it’s still a game with odds. It would be possible to build a product that specifically avoids lock-in, but what kind of business would that be?

Finally, let’s consider the security risks.

Evernote – the app, the service, the company – wants to be used for as much of your information-world as possible. I wouldn’t be surprised if they build in e-mail at some point. And just as with e-mail, the result is that a lot of your very private information lives, unencrypted, on somebody else’s server. Or in somebody else’s “Cloud.”

That doesn’t mean it necessarily will be compromised, but it does mean it can be compromised. Through your own stupidity or theirs. And as Evernote grows, it becomes a target for government-sponsored cybercriminals who have beaten the likes of Google, so it’s not just a questing of hiring good programmers. Evernote’s security was breached in 2013, and it probably will be again one day.

This problem, much like the lock-in problem, exists with any useful service, because the things that make it useful are precisely the things that limit its security.

In this day and age, sadly, you should assume that everything stored outside your own (encrypted!) hard drive will eventually be stolen by Bad Guys. Hopefully it won’t, but consider how private you want it to remain if that should happen.

Again this is something you could design for, but it’s not clear that would be smart business. Security always comes at the expense of convenience.

I don’t think “normal people” will ever really care about lock-in or security, and Facebook is the standard Exhibit A of this argument. I don’t think Evernote needs to lose any more sleep over these things than they already do. As the man said: If you’re not cop you’re little people.

But I, in this metaphor stew, am cop.

A Roll of One’s Own

I’ve visited and revisited this problem so many times over the years, and cobbled together so many half-assed “solutions,” that the temptation is very great indeed to just go all-in with Evernote and let the Everelves manage my Everstuff for Ever and Ever.

Yet still, there is the classic hacker’s weakness: I am always tempted to roll my own. (Back when I was a smoker, I rolled with the best of ’em. Just FYI.)

If I were to build a Note Manager Thing, its goals would be simplicity and security.

Simplicity would mean a strong bias towards plain text as the source format for, well, everything; and human-readability in addition to machine-readability wherever possible in metadata, i.e. no XML ever; and the assumption that anything else manipulating the notes is probably operating in a UNIX-like environment. These ideas are, unsurprisingly, very close to the original rules of Markdown.

Security would mean that everything is encrypted such that only the user can decrypt it. That’s a hard nut to crack of course. If I’m editing a bunch of text files in folders on my Mac then I can just stick them in an encrypted disk image et voilĂ , encryption. But then there’s the problem of moving things around, and of doing cool machine-learning tricks on the server, and a million other things. In reality I’d probably have to encrypt each piece of data separately, thus making the software slower and introducing some nontrivial number of bugs. What about auto-save? Ad nauseam.

Oh yes, and usability. That would be a goal too, at some point.

Just because it’s hard doesn’t mean it’s not worth doing.

It doesn’t seem quite sane to imagine this working as a business, at least absent some Silicon Valley Venture-Capital Slush Fund. Maybe if I called it

Just because it’s not a business doesn’t mean it’s not worth doing.

Furthermore, I have a lot of things going on. I have a job, I have a wife, I have a gym, I have a painting studio (sort of), I have a photo lab, I have half a dozen experimental software projects, and I have a taste for the good life.

So for the moment, for me, it’s probably not worth doing. I’ll probably continue to use a bunch of different programs, including Evernote, for my digital writing and writing-down. And I will probably return to think about the problem once or twice a year, secretly hoping someone who shares my design goals has gone and built the thing I would build so I can just trade him (or her!) a few Merkels… er, € for it.

On the other hand:

If you’re listening, Marc Andreessen, I wouldn’t say no to your money.