Hacker News Reader

I fully agree, but as a vendor that only implements bring-your-own-storage, I can tell you it's not much fun from the supplier standpoint. Our no 1 support request is something along the lines of:

"I've deleted my diagram, can you restore it for me?"

"No. We don't store your data, we never even see it"

"What a useless service, just get my data back or I want to speak to random escalation person"

99% of users don't care about privacy, architecture, etc. They want it to just work (TM). They want us to be able to undo their deletion, find their data when they forget where they moved it.

So it's users pressuring suppliers in the wrong direction, it's not all one way traffic.

> So it's users pressuring suppliers

So may be managing data by product supplier is NOT the wrong direction?

It's about expectations. Users expect supplier to store and manage their data. If bring-your-own-storage became mainstream, users would know that it's them that manage their data and wouldn't have these expectations on the supplier.

(Disclaimer: I work at Google, not speaking for the company, etc)

We can do both. Google stores, backs up, secures, and generally ensures your data is safe, and you can click a button to download it all if you think you can do any of that better.

This is probably too hard to ask every company to do the same, as it would exclude smaller companies who don't have the resources (this takes a lot of time and development!). So it makes sense for them to offer a different feature set, and accept that it's not going to be suitable for some people.

Just for info, the last time we ran stats on which cloud storage our users pick it was:

Google Drive 53%

Dropbox 4%

OneDrive 1%

I don't think you're having a problem selling the idea :).

And where did you got that data from? Because where I live dropbox is far and ahead in user-base compared to Google Drive. That's the feeling I'm getting, mind you, I don't have any actual data... But even the apps I (2do, 1Passwd, etc.) use have support for Dropbox while none of them has support for GDrive.

I got that data from our SaaS application that offers storage using those 3 providers. Our target market is much more business over consumers, I suspect the figures look different in the consumer market.

Oh cool.

> We can do both. Google stores, backs up, secures, and generally ensures your data is safe, and you can click a button to download it all if you think you can do any of that better.

It doesn't really secure users' data, since Google is able to read that data. Were it truly secure, it would all be encrypted with a high-entropy keys known only to the users.

People without backups could always choose to store their keys with Google, but people who want their data to actually be secure would be free not to.

> This is probably too hard to ask every company to do the same, as it would exclude smaller companies who don't have the resources (this takes a lot of time and development!)

What's hard about "Backup to google drive / dropbox / icloud"?

I've seen single-developer applications do this. If a company of 10 people can't do this, they need to question the competence / diligence of their team.

> We can do both. Google stores, backs up, secures, and generally ensures your data is safe, and you can click a button to download it all if you think you can do any of that better.

This is exactly the right solution, and credit to Google for providing it.

> This is probably too hard to ask every company to do the same, as it would exclude smaller companies who don't have the resources (this takes a lot of time and development!).

I'm surprised you say that; it doesn't seem to me that 'dump this user's dataset to JSON/XML/whatever' should be very hard to do. Admittedly I've never worked on anything at Google scale. What am I missing?

> What am I missing?

To pick a few things I can talk about - sufficient replication to not lose data, data security that actually works (small companies can't - and shouldn't - afford disk shredders and internal access audits), targeted obliteration of customer data on request (including getting it out of the backups without rendering the backups worthless), and making sure that your exported data dump which contains massive amounts of personal information is only ever released to the correct user.

Doing any one of these things in isolation is not all that hard, because you can trade off against the others (for example, it is dead easy to do a data dump if you don't care about data security). Doing all of them at the same time is a large, expensive development project that would sink any startup.

Users no more expect vendors to lose their data than they expect banks to lose their money. I suspect we'll eventually end up with some sort of standards body making sure you do backups.

I don't quite follow your point. We didn't lose their data, we never had it.

I've probably put the comment in the wrong place in the tree, sorry.

The analogy then is having the bank tell you to take your deposit home and hide it under your mattress. That mental model just doesn't fit into people's world-view.

> 99% of users don't care about privacy, architecture, etc. They want it to just work (TM).

This is the sad Catch 22 we live in where privacy is jeopardized at the cost of ease of use

Do you think there is any value in educating the customer further? Periodically (5mins) reminding the user with a small notification to save (and backup) their work? I agree it can become a UX Nightmare (e.g. MS Paperclipguy), but having a checkbox "do not remind me again" could possibly go a long way as well.

We autosave to their cloud storage. The case I'm talking about is where they manually delete files on their cloud storage and expect us to restore them because we're the app vendor that originally wrote the files.

Yeap, it's a 1990s model. I had a discussion with a software salesman... I was explaining to him that, what is valuable is not the software it's the data I will have put in his software in 5 years from now and if I don't have raw access to this data, I don't want his software.

He didn't understand a word.

Oh he most certainly did. Companies don't want to give you data for exactly that reason - if they can convince you to start using their software, you'll quicky be stuck with them.

At the moment, people are implicitly trading their personal data for services and convenience (communication, transportation, etc). Yet our data has value not just aggregated across millions of users and trained to sell ads; it has value to us as individuals: third party platforms can disappear, get hacked, act dishonestly, or change their business model.

There have been advances in systems that reclaim or duplicate ownership of third-party data back toward the user, personal API/quantified being the most recent. I think the missing link is not the technology, but rather a social movement that promotes the value of data access and ownership. This might be one silver lining to recent data breach and hacking scandals.

or is the problem just a poorly designed data model by Evernote? So for instance, i have had a github account since 2009, and i access both my personal repos and my repos that belong to my employer using the same accountName. When i joined my current employer, i no longer had access to my prior employer's repos on github, though my access to my own repos (both public and private) remained unchanged. And when they switched from github organizations to hosted github (more similar to the OP's situation) all of my repos in my personal github account remained intact and my access wasn't affected in any way

Sure, GitHub is better designed, but it also follows the principle of data freedom espoused by twhb and nsrivast. That may have come naturally by virtue of Git, but they also provide an API with no restrictions that can be used to move other things like GitHub Issues.

@nsrivast: a movement is exactly what we need. EFF is great, but they have a lot of things on their plate, and are not always consumer friendly. I've been working on core ideas for a Users First movement, because the web, contrary to the ideal results of a free-market, has been putting users last. Much of that has to do with users being the product not the customer. Your ability to "organize & visualize information for efficient teaching & understanding" would be invaluable!

@twhb: you have no email address on your account. please contact me if you are interested in helping articulate such a movement.

I am interested.

Github took the intelligent approach of separating identity from authorization. You have your own identity, which doesn't change. On top of that, you have a set of authorization permissions which can be updated, added to, and deleted from. Changing authorization permission X doesn't affect authorization permission Y.

Evernote seems to instead have lumped identity and authorization together. And had only one set of authorization permissions: all of your stuff. So when the permissions change, they all change.

This approach is nothing more than laziness. It's not just "not thinking ahead", it's completely ignoring 40 years of practice in the field.

Although you can't browse it, you can instruct Evernote to keep a local copy of all your data, allowing you to keep incremental backups. You can find (and change, if you want) the database folder under Tools/Options, General.

I do agree with your other points, although there are significant limitations to using the filesystem for storage in an application like Evernote. Perhaps a good compromise would be to use a publicly available database engine like postgres (although I'm sure they derive advantages from their proprietary format beyond the mere fact that it's proprietary).

This is reason #1 why I like DEVONthink for the Mac. It uses exactly this model. I'm sure equivalents exist on the other platforms, but I'm not all that familiar.

I have been an Evernote Premium user for a long time, and since the company seems to be having a rough time, I recently looked at getting all of my data out of Evernote. What a nightmare.

I loved Evernote. I loved the fact that Phil Libin said he wanted it to be a 100-year company. I don't know all the details, but the fact that he seems to have given up... what a letdown. I'm working on getting my data out. Once I'm done, I think I'm going to let my subscription lapse.

I use DEVONthink too for the same reason. The only thing DEVONthink lacks is cross-platform support. It's a pity that it's not available on Linux or Android.

To be fair, it lacks a lot more than that when compared to Evernote. Evernote is a bundle of great ideas, done horribly catastrophically badly.

1. Evernote has dual-language OCR of PDFs and images (and a few others crappy formats that normally shouldn't be used like .docx, .xlsx, etc). That's a great idea. DevonThink doesn't even have Japanese support (their licensed OCR only supports Western languages).

2. Uh... OK god dammit that is really the only feature that keeps me using Evernote's horrid apps. Shit.

OneNote has 0CR from images. It works fairly well.

At risk of sounding like an Evernote apologist, I'm going to repeat a question I asked to another commenter elsewhere in this thread:

Why do you consider it a "nightmare" to get your data out of Evernote? You can select one, several, or all of the notes in a notebook and export them in either HTML or XML format, with all metadata (tags, etc.) and embedded media included. That seems pretty good to me, and it's certainly easy. If you wanted something more sophisticated, Evernote on the Mac has exceptionally complete support for AppleScript, so you can code up whatever kind of fancy extraction you need.

I'm not asking to be argumentative; I've always considered Evernote's support for extracting data to be pretty darn good, and clearly superior to many competitive apps/services, so I'm curious to know what additional functionality you would require before you'd say it's not a "nightmare" to extract your information.

If I put plain text into a note, I'd expect to get plain text out. You put plain text in, and you get something that still have to extract the text from. That isn't what I'd consider "pretty good".

My point doesn't apply so must just the plain text. It's everything. You put it in in source format, and you get .enex out. It's not so cut and dry.

Going in, it's one step. Coming out, to get it back in the same format you got it... it's multiple steps, and manipulation with other tools.

This is a great point. There's an analagous situation in finance: would you rather have a single mutual fund own all your investment money, or do you want a brokerage account with some active management? The latter is definitely more complicated but gives the user more control, accountability, visibility, etc. While it makes sense for a Whole Product to want to hide this complexity and liquidity from the user, any long-term user will very likely want it after they've learned to use the product.

Actually I'd say a large percentage of people would be very well served by a globally diversified index fund for all their equity holdings...

And perhaps the analogy is still useful. I'm ok with that situation because I understand and accept the logic of indexing. If the fund I owned changed their prospectus and decided to start active management, or significantly increased their management fee, or whatever, I could pull my money out and invest it elsewhere. Similarly in software there are definite advantages to having a dedicated steward for a given type of data, but it's much better if there is some way to take your data back should you decide to put it elsewhere.

With Evernote, it is possible to export all your notes in a variety of formats, including XML, which can then be imported into other note taking applications if you wish. So to me, the bare fact that the application controls the data isn't necessarily an issue. What was becoming an issue was the fact that they didn't appear to treat it with sufficient respect. I hope we'll see that start to change now.

Thus... I use https://github.com/Jermolene/TiddlyWiki5 for the most vital information. It's more tedious, local (a simple file) and run in a browser.

> is a principle fundamental to Evernote: data owned by an application, rather than data accessed by an application.

Close, its a principle fundamental to modern cloud-based companies.

Offer the cloud storage for free, for 0 privacy and the permission to read, copy, access, and sell your data to third parties

> Close, its a principle fundamental to modern butt-based companies.

> Offer my butt storage for free, for 0 privacy and the permission to read, copy, access, and sell your data to third parties

Moreover, data loss is not a possibility, it's pretty much a certainty, because it's backed straight in the business model. At some point investors need to get their money back, and what better way to do that than have the company acquired? It's such an effective way that a lot of people starting their companies do that hoping that they'll get bought by a big corp. That the product gets subsequently shut down doesn't seem to bother anyone (except the customers, who at this point really should know better than to do business with a startup unless they're prepared for the moment said startup shows them the middle finger and disappears).

Completely agree with this and I think the model of an application that accesses data is the only one I will trust going forward. Great example of this is Scrivener. I can have my files locally stored or in the cloud, but Scrivener just accesses, not owns the files.

And this is why iTunes is a much better software than iPhoto.

>I sound like a cynic when I say this, but the root cause of failures like this is a principle fundamental to Evernote: data owned by an application, rather than data accessed by an application.

No, you don't sound cynical at all. I think It's perfectly reasonable to think that and it seems to be shaking out that way. It does seem pretty douchey to pontificate about your views on applications and data access in the form of a "this is why I don't use applications of this type" and then launch into an example of exactly what happened to someone who just said "I lost all my data with this service." It's kind of like "no shit. Thanks for telling me after the fact, professor."

> root cause of failures like this is a principle fundamental to Evernote: data owned by an application, rather than data accessed by an application

Not really. It's not like MS Word reading your .doc find can't corrupt it leaving you hanging if you don't have a backup. Catastrophic data loss isn't cured by the "data accessed by an application" model.

At least you can backup files. Applications, not so much... I'd add that data should be stored in a standard open format, so that an application failure won't prevent you to 1. keep your data (because backups) and 2. use your data (because open format)

It's better because you can back things up, but people tend to have a bad track record on backing things up unless it's part of their job description.

> an application failure won't prevent you to 1. keep your data (because backups)

My point is that you don't inherently get backups just by having control of the data file. It requires extra action on your part. If you don't backup, an application failure can easily destroy your only copy.