Hacker News Reader

I think the poor architecture is a direct result of there being too many features.

Evernote has shoved so many features into the product that they haven't been able to sufficiently hone individual features. This leaves them open to competitors who focus on a subset of features but do a much better job of implementing those features.

Apple and Google get around this in different ways. While Apple has a broad set of products and features, they're pretty conservative in their pace of releasing new features. So they have plenty of time to hone and integrate new ones.

Google seems to have solved this through radical decentralization. They don't really require that different products form an integrated whole. This allows individual teams to focus on delivering high-quality products without the architectural overhead of bundling everything into one product.

Perhaps ironically, Evernote alienated thousands of power users when they did their big 2.2-->3.0 rewrite a few years ago. Evernote 2.2 on desktop (+ the browser extension) was tremendously awesome, and then they rewrote the whole thing in .Net and basically made the desktop client a webview, in the process removing about 90% of the features people loved while causing all the sync and stability problems people are still complaining about.

I want to love Evernote because having a centralized, cross-platform, easily searchable repository for both binary content and text is hugely powerful, and there aren't really any viable competitors in this space. But, I don't trust the company and I don't trust the product, so I won't use it, and I'll stick with several different other products that together mostly fill the same use cases for me. One of those is Google Drive, but it's even clunkier than Evernote (although sync works great!).

They fell into the big rewrite trap ... possibly driven by developers hoping to clear out some cruft

http://www.joelonsoftware.com/articles/fog0000000069.html

The article is dubious at best. You hear about the big failures or notice when rewrite releases take forever, but you never hear about the companies that slowly lose opportunities or fail to keep up with innovators because they keep trying patch a crappy codebase. Nor do you necessarily hear about the successful rewrites since they just happen under the covers.

One of the major points of the Joel article is that the best option is usually one you didn't name; incrementally fix the "crappy" codebase. At no point do you do a "big rewrite", at no point do you have a big step back, at no point do you lose the ability to make forward progress because the new code isn't ready and the old code is deprecated, etc. Even if it may take somewhat longer to get there, the integral of value over time often still comes out larger for incremental improvement.

Developers want the default answer to "abandon the old mess and write a new one" (snarkiness fully on purpose); Joel's point is that the default answer ought to be incremental improvement. Not that it's always the right answer, but other answers ought to be scrutinized more closely than developers might like.

From a professional point of view, it's actually perfectly fair to consider that greenfielding a new project with hot new tech (or even "newer"-but-established tech, my personal favorite choice) is more fun than trudging through old code. Human factors matter a lot. But we are also professionally obligated not to overprivilege it.

Besides, if you treat it as a serious project instead of a series of hack jobs, in my experience, very serious incremental improvement still offers a lot of engineering challenge and fun. I think one of the biggest mistakes people make is to prejudge incremental improvements as a hack job, when it becomes a self-fulfilling prophecy.

Incrementally fixing the "crappy" codebase is great in some cases: Michael Feathers's book on rescuing legacy code is fantastic. However, the limit comes at the language barrier.

Consider a COBOL codebase on a mainframe where you cannot find developers interested in learning the language and developing against the mainframe is convoluted -- you may not be able to pay talented people enough to toil in those coal mines, and you have to have to overpay subpar talent (or chase after the handful of expert mainframe COBOL developers.)

You can incrementally rewrite, just like you can incrementally refactor. Stick an API gateway in front of your COBOL mainframe that responds the same way it does, and then stand up a new well-architected service, microservice by microservice, that has a good API—and have the API gateway query the new service (using its nice API) whenever clients make calls to what they think is the legacy service, passing whatever calls you haven't re-implemented yet through to the legacy service.

Eventually, everything will be on the new services, and you can shut down the legacy COBOL system and just keep the API gateway there to pretend. (If you can get clients switched over to consuming the new APIs directly, you can shut down the API gateway too—but good luck with that; their side probably has mainframes too.)

It's called the strangler pattern by many. (You probably know that, but our dear reader may want to follow up.)

It also assumes that you have a proper API to start. It assumes that you have the organizational maturity to handle synchronizing two separate systems and the distributed transactions that entails.

The other problem not mentioned in the rewrite/refactor conversation is that the most common reason for rewrites is that the business has backed itself into a corner and the assumptions under the first system do not apply to where the business wants to go.

"It assumes that you have the organizational maturity to handle synchronizing two separate systems and the distributed transactions that entails."

Well, to be honest, when we're talking about proper maintenance and advancement techniques, we must by definition be discussing the topic only for those with the discipline to correctly implement relevant techniques and policies. If your developers or management choose not to, be it for whatever reason up to and including total lack of requisite talent or experience somewhere, you've already lost and the only thing that can possibly save you is to address that problem first, and if that's not possible for whatever reason, you've simply already lost.

As a result, it turns out not to be an interesting case to discuss. Even if it is, probably, the dominant case in the field....

I think it makes the most interesting case. How do you level up an organization? It's a very hard problem that hasn't been solved yet.

In some markets, the talent to handle this type of project doesn't exist. In other markets, the cost to acquire that talent is more than the marginal cost of rewriting the system.

What you are saying sounds cavalier to me. Fact is, big rewrites are very risky and there's nothing appealing about working on them, I can attest! Any developer worth his salt should fight tooth and nail to kill such a rewrite project and then run screaming when business and management go forward anyhow with scrapping the working legacy system in order to "start fresh".

Lost opportunity does suck. It would be nice if the first system was written in an extensible way with lots of tests and correct documentation, but it rarely is. All software needs to be replaced at some point. All software has a finite useful lifespan. So, this is a common problem. But you must have a responsible plan for replacing any first system. Any sort of project plan that has a magic day where the legacy system is shutdown and the second system is turned on is reckless and absurd. I've heard of failure rates for such projects being 80%, with the rest going way over budget before completion.

The best you can do when replacing a crufty old system is to employ strangulation. You slowly, methodically kill off the legacy system one feature at a time with the equivalent features in the well-written/tested new system. You run both systems side by side, sharing the processing load until one day, years in the future, when you have replaced all the features and the legacy system has nothing to do. It's a hassle and it requires everyone to be organized from start to finish, but you eliminate the big-bang release risks.

Luckily, with all the web services and factored out componentry we've been building in software over the past 10 years, the strangulation approach has never been more feasible.

There's lots of good youtube lectures and agile community blog posts on the topic of second systems and strategies to implement the strangulation. Here's an old classic that introduces the concept better than I could:

http://www.martinfowler.com/bliki/StranglerApplication.html

Or perhaps by the chasing the universal / web view on the desktop unicorn. Native apps still best HTML frames on the desktop in terms of functionality, and no amount of cost savings for the vendor will convince users otherwise.

Multiple code bases may be a pain, but they still produce better UIs when you support more than one OS.

> Multiple code bases may be a pain, but they still produce better UIs when you support more than one OS.

I don't know if that's universally true.

Slack uses Electron for its desktop apps, yet has received a lot of acclaim for its UI and seems to be trouncing alternatives with truly native offerings (IRC, HipChat).

I have always been surprised they get so much UI praise. To me, it feels like the HTML frame that it is, and often confounds my expectations of how a native app works. It is definitely better than its competitors, but that doesn't mean it's all that great. Having said that, it seems like they've wielded the web view approach to achieve a great flexibility and development speed advantage, so it's definitely a valid way of doing things. But I keep waiting for them to switch to full-native now that they're well established and can afford way more developers.

I'd hardly call slack the epitome of usability. It guzzles memory, is laggy on my workstation (Xeon cpu lots of ram and a high end gaming gpu), their search functionality is abysmal too. I quite like slack, but I really wish they would consider a native app, or even just address the performance problems of their chat client.

I do not really understand how Hipchat can be so bad though. The client is ghastly on all platforms and not being able to handle shaky internet connections these days is a joke.

Slack has 2.8 billion reasons why you're absolutely, painfully wrong about this.

Apples and oranges.

Slack's story is more about their service and 3rd party integrations, less so their client apps. A messaging client needs far fewer client features than a service like Evernote, and can get away with a simpler development environment. A web frame in a box works for Slack, but not for Evernote.

Evernote's client from a feature and polish perspective is as important as their service. The service need to store and sync (where they also have problems), but the client has to cover a multitude of other features. Cutting corners on the client layer to ease development has had an obvious impact on the UX.

I believe they made about 30M in 2015. At an evaluation almost 100 times revenue, I think this story isn't done playing out yet.

I don't understand what you're even trying to say.. aren't companies expected to ramp up their revenue over time?

Sure, but a 100x revenue evaluation is incredibly exuberant. For reference, I think Facebook hovers around 15x revenue to market cap.

Yes, because the best way to make software is to eternally add more features to the same pile of code until it is impossible to get anything done.

When my room is dirty, I clean it. When it's really dirty, I pay a professional to clean it. I don't burn my house down and start over.

And when your room is so dirty that there are bugs in the walls, termites eating the wood, the thermostat is leaking mercury, the ceiling is dropping lead paint + asbestos, and the foundation is cracked and pulling the house apart ...

you burn your house down, and start over.

As technology people, I agree that we lean far too heavily towards the "rewrite it" option than we should. But that doesn't mean it's never the right option.

Worse than that - if just the foundations are bad, you may have to start the house over. Even if the house is fine but you get the wrong type of mushroom in the wall of 1 room, the house will have to be rebuild.

Strange analogy by GP, you have way more possibilities of fixing stuff in the software world than in the physical one.

I deeply get your point, but to step away from the metaphor, usually the problem of shitty code is a lack of focus on continual refinement instead of adding features, which is a management/process problem. If you have a serious problem with your process, you're never going to build a solid code base.

Hell, at my last company, we have a big server product made of duct tape and spit, and about halfway through a total rebuild, we had another ball of spit and duct tape made of a slightly trendier stack. We added coders, we brought in front end guys, but guess who we didn't replace?

>want to love Evernote because having a centralized, cross-platform, easily searchable repository for both binary content and text is hugely powerful, and there aren't really any viable competitors in this space.

OneNote?

OneNote is the best current option, by far, but taking advantage of a lot of its best features require also using MS Office & Outlook, which is a nonstarter.

> Evernote has shoved so many features into the product that they haven't been able to sufficiently hone individual features.

...and that's a really hard problem to fix as you need management (who may not understand the root cause at all) to be on the same page as engineers working on the lowest level of the product.

If only they had some sort of automated communication mechanism that they could use to share the same pages of information between the engineers working on the lowest level of the product, and the management that needs to understand the root cause of it all...

Only if synchronization is ad hoc. The world will no longer tolerate throwing another switch and gauge on panel and calling it good.

Hear hear. What finally prompted me to switch from Evernote to Workflowy (even though Workflowy has maybe 1% of the features that Evernote does) was that when I opened the Workflowy app on my phone, I could be assured that I would see exactly what I'd finished entering on my computer, even if I'd only entered it a few seconds ago. Heck, I can enter stuff on my computer and watch my phone update in almost real time.

When I first saw Workflowy's synchronization, coming from Evernote, it seemed almost magical.

> Apple is mentioned as another company with many products - but the overwhelming majority of both specific products and specific features work really well.

Really I find their software to be hazardous to my health. I am okay with people liking their hardware but their software and services makes me frustrated 90% of the time.

> their software and services makes me frustrated 90% of the time.

And the other 10% is spent in Keynote.

Seriously, Keynote is excellent, but Mail is the only other Apple-made app I use, and that's only because I haven't found a better Mac mail client that works with Exchange.

Apple Computer's "Lemonade Stand" [1] was pretty good educational software for learning how to run a small business, while mitigating the consequences of global warming.

[1] https://www.youtube.com/watch?v=0dUbDDmXp2I

http://airmailapp.com/

I’ve been using it for years now, and it is fantastic. I do use the beta versions, but I’ve experience very few problems.

Outlook 2016? It's a huge improvement over previous Outlook iterations for OS X.

> The problem is the crap architecture behind its synchronization model that leads to a sub-par user experience.

One 'bug' that's been present since I started using Evernote: Create a new note, paste content, start typing a title, pause, and have it wipe out the title with part of the content or move your cursor to the start of the title.

What you mention indeed is a problem for its core users who already use Evernote. However "too many features"--each of which is mediocre--is a much larger problem. I remember trying to use and like Evernote multiple times through its evolution and not really sticking because there wasn't a single feature that was uniquely superior. Why use Evernote for saving to read later when Instapaper or Pocket do much better job? Why use Evernote to keep todo lists when there are tons of other Todo specific apps that do much better job? etc.

We could say the CAUSE of the problem is too many features, because makes them unable to solve the most important problems.

I have never experienced the problem you describe, but the lack of decent search functionality bothers me.

Have you tried Onenote ? How does it compare ?