Hacker News Reader

This makes no sense. The actual act of programming is completely agnostic to what the bytes you're processing mean.

They could be the most boring data in the world, but if you're trying to juggle enough of them at once, at high reliability, you're going to end up with very deep technical challenges.

I'm sure somebody probably said "why would you want to write code for an online book store? That's so technically shallow and simple." And then we got Amazon.

Great hackers redefine the problem in front of them, and that act of redefinition is the interesting part. No program is interesting if you're simply translating a flawless spec into code. Any program is interesting when it's entirely up to you what the spec should truly be.

There are good and bad programming tasks in all fields. But the kind of "complex simple software" the GP talks about, where the actual task is mind-bogglingly simple but there are hundreds of edge cases in the requirements, is a lot less fun to write than something with more cohesive requirements (i.e. where you don't have to match what an existing business does)

there's a challenge in how do you structure and organize edge-case logic such that the core application design needn't be entirely redesigned every time a new requirement is revealed or an existing one needs to be altered. When the client comes to you with changes in requirements, having built an architecture that can turn on a dime is a rewarding experience.

It is definitely true. If you manage to build abstractions that don't leak in the wrong places, it'll welcome changing requirements very well.

I buy into the ethos that HN is an echo chamber and, in private, many here have agreed with me. In fact I'm almost certain the reason I was forced to create a new account here after over a year in good standing is because I rejected a handful of this echo chamber's views. That said, I still value HN and like it but have learned to filter out a lot of the bad stuff (which keeps on coming faster and harder lately).

I also think a lot of what you said can be applied to the B2C space:

B2C is is boring as shit and not even remotely interesting.

The only challenge to setting up a competitor to Facebook/Twitter is that it's so mind numbingly dull. B2C programming is an utter pile of mind-numbingly boring crap. It's not sexy because it's incredibly simple with a colossal amount of edge cases. B2C is mindbogglingly complex simple software.

B2C is just the edge cases.

It's not the remotest bit challenging to most people. It often lives and dies by the latest fad and undeserved hype in tech blogs.

But you'll get a lot of VC funding.

Do you want to read about it? No, because it's so ridiculously shallow and played out. That's why everyone's sick of it.

(disclosure: I'm actually writing B2C software (no, really, I seriously am, not joking or being sarcastic at all there))

Sexy is in the eye of the beholder. If you characterize all B2B software like Quickbooks then yeah, it'll be boring to most people in the same way that yet another social network, geo-whatever, so-lo-co-mobile app is going to be boring to most people. You're right about the edge cases when you label them as edge cases. But if you think of them as niches then it becomes interesting again and a whole new market is laid out before your eyes.

Your whole comment was very general and you seem to apply your personal opinion to "most" people. I do get what you're saying and I can see how a lot of B2B products can be boring to a lot of people but the way you put it makes your argument weak.

You're defining B2B too narrowly. You describe what sounds like a business logic app, probably mostly CRUD. And yes, that's pretty boring. But that's not the only software that businesses buy. Heroku, for example, is a B2B company because they sell predominantly to companies, not individuals. Heroku might not be something that the founders' relatives and college buddies would use, but it's far from boring.

I think there are two kinds of B2B startups that can be super interesting:

The first is eliminating entire classes of jobs and the second one is eliminating entire "glue industries"/middlemen (disintermediation).

The former may sound callous, but creating a solution that frees up a human so they can pursue greater endeavors is a noble goal.

Those are exciting problems.

I think there is a big difference between B2B and enterprise software. Enterprise software, to me, seems very boring. B2B software like, say, the stuff 37signals makes is interesting. I enjoy those business models and reading about b2b web apps and picking apart their UI designs and how they go about solving problems. It seems like it would be a ton of fun to build something like mailchimp.com or basecamp.com, etc..

Agreed. My goal used to be to create something that my friends would use. I hated "enterprise" software. Thought it was bloated and ugly. Then I saw 37s. I grew tired of B2C software. Now I'm working on software I want all my peers to use. My friends wouldn't get value from it.

But, I can tell you it is the most fun I've ever had building software. We get to choose our technology. And we're using technology I consider sexy - like Ember.js and some great database technologies that make the hard stuff our app is doing ridiculously fast.

I get excited about things like making our queue tremendously fast by switching from resque to sidekiq.

There's nothing I'd rather be doing than building B2B software.

So Github is boring? 37Signals is boring? Heroku is boring?, Photoshop, Maya, LogicPro is boring?

There sure are boring things in B2B just as there are in B2C, but there are plenty of B2B business that are great fun and challenging as hell.

Don't forget RethinkDB, Mixpanel and Cloudera - those guys are solving some difficult, real world problems. Personally, I wouldn't classify massively distributed computation as a service boring. If anything, the scale (in numbers of users, amount of data) makes B2B of this kind even sexier than share-my-cat-at-my-location-agram.

Hmm, I get to write radio communication software, mostly TETRA stuff. Playing around with lots of awesome hardware, writing a lot of both high level management software and low level bare metal code, with many _very_ non-trivial problems. I find this incredibly interesting.

There is such a massive load of cognitive dissonance about B2B being sexy in the responses you got. Is it really such a blow to one's ego to admit that it's boring but profitable? Yes, there are technical challenges, no, the majority of it is not interesting.

Setting up a challenger to QuickBooks does sound dull.

Setting up a challenger to SAP on the other hand does not sound dull.

Wanna see a cool B2B company? Check out a tool called CyberArk.

> Setting up a challenger to QuickBooks does sound dull.

FWIW I find Freshbooks to be a pretty exciting product as a user. I'm sure I would enjoy working on it as a developer, at least for a while.

Can you really think of nothing more interesting than accounting software. You lack imagination.

The only challenge to setting up the ultimate competitor to Sage/Quickbooks is that it's so mind numbingly dull.

Yes, the dullness of the topic is why no one has bothered to enter the market and compete with the $3 billion Intuit, or the $1.5 billion Sage. It's just too boring for them to bother with that kind of money.

You're fing kidding right?

Do you think that maybe, just maybe, the 17,000 pages in GAAP have something to do with it? Or maybe the 3000+ counties.. or the long list of county/city/state/federal rules and regulations? and not just accounting rules, but employment taxes, credit card processing, invoice generation, etc, etc, etc.

I think you don't understand the complexity of the topic and don't know the full feature-set of sage/quickbooks, and that's why you think it's simple software.

I think he understands it rather well - it's "complex simple software".

You can think of any software as consisting of a set of problems of varying degree of "difficulty". Sort them in ascending order and plot them. This graph will give a very tell-tale signature as to what you're facing.

B2B / enterprise / call it what you will software has an infinite number of individually easy but collectively mind-numbingly painful problems. Yes, a lot of that bs comes from laws/rules created by "others".

"Interesting" software has a small-ish number of very difficult problems and then a manageable number of easy ones. These are the type of problems where you can take a small team of very smart people and really knock it out of the park. They are most unfortunately very few and far between.

Social cat photo software I can't speak to. I'm certainly not implying that it's in the "interesting" bucket...

No, he definitely doesn't understand what quickbooks is. He says 10million loc is an indication it has gone "horribly wrong". That tells me he doesn't have the vaguest idea what the feature set of quickbooks is.

I've been programming for 15 years, and have personally used sage for 7 years. 10m loc is definitely within reason for the number of features the app has.

If he doesn't understand WHAT the software IS. Then how can he sit there and say what the level of difficulty is.

He DOES NOT KNOW. because HE DOES NOT KNOW what quickbooks is.

I parsed his "something's gone horribly wrong" as to mean "wrong with the accounting laws in this country", not "wrong with the software development".

Law is complex because the world is complex. We're talking about "essential" complexity; it exists in the domain and it can't be wished away.

I have to disagree. Law is complex largely because politicians need to feel like they are doing something. They make changes so they can claim in their next campaign that they've improved things. They make changes to benefit their likely donors/voters. The cruft keeps piling up and it rarely gets wiped away.

Consider tax-advantaged retirement savings. Why do we have: traditional IRA, Roth IRA, SEP IRA, SIMPLE IRA, 401k, Roth 401k, and 403b, all with different requirements and different benefits? The need to have a reasonable amount of savings when you retire is universal, so why this complicated mess? Why are two people doing the same job at two different companies able to save dramatically different amounts based on whether or not their employer chooses to offer a 401k? Why force every taxpayer to do a 10-line (or whatever it is) calculation to determine eligibility to for an IRA just so we can prohibit a few rich people from having a benefit (hint: it would be a lot easier for everyone to make it universally available and just bump the rich people's tax bracket by a hair if you really feel the need to wipe out the pittance in tax savings that they would obtain). This isn't necessary complexity, it is pointless stupidity that we all suffer for (except for the accountants that gain employment from it).

Sure, that's one source of complexity in legislation.

But really, seriously, the world is complex. Every day people stumble on combinations of circumstances that have never occurred before in human history. Some fraction of those people get into an argument about what happens next. Some fraction of those go to a judge and ask her what the answer is.

The judge takes the existing principles, cogitates a bit, then extends the case law to cover the new scenario based on analogy with older scenarios.

This creates a new piece of law. It's never been seen before. But it might turn out to have profound consequences.

Before switching to computer science, I studied law. It's complicated because people are complicated.

Most people are thinking too narrowly of B2B, and while it may be true that business logic and accounting are dominating in number of installations / volume of sales / whatever, some very unsexy businesses (road toll collection, what could be worse!) do want most sexy science and technology. Significant fraction of computer vision studies is dedicated to vehicle recognition and classification.

I think ther are some sexy B2B problems, ie problems that can solved with technology rather than business solutions that use a tiny bit of software.

Analytics is pretty huge. Companies will pay big dollars for profitable insights into their data, queue distributed computing and machine learning. In machine learning and AI there are a massive number tasks being done by unskilled temps that could be automated away, for example data entry, basic research, report interpretation (back to analytics).

Then there's whole field of taking almost any piece of enterprise software and just giving it a decent/usable user experience, using sexy Web tech, html5, etc.

I think the only thing unsexy about b2b is that you'll never be famous in your friends eyes, ie you'll be the guy who made their accounting software 30% less frustrating rather than the guy who created the latest local cat photo montage social network