Ask HN: What do you hate using but are forced to use?

SwellJoe · on Aug 2, 2009

So, I wonder if you're thinking of starting a business based on one of these "I hate using" stories. If you are, I'd advise caution entering a market you don't actually understand personally. Yes, we all make software that eventually grows to do things we don't use ourselves (I use functionality that touches less than 10% of our total codebase on a daily basis, and never hit more than 20% of it, for example), but without understanding the core problem set of your target audience, I'm not sure you can build effective software to solve those problems.

When I worked in scientific computing, one of the founders of the company I worked for had done doctoral work in the areas the software was covering, and the company hired domain experts, who also happened to know a bit about software development, whenever they moved into new fields.

This is the reason I've never gone into medical billing, despite the fact that I know the medical billing software industry is a mess, and also an incredibly lucrative market. I know it is a mess, because my sister owns a medical billing firm, and I've seen the software her company uses. But, without spending a few months working in the field, I wouldn't be effective at solving those problems. If I ever did decide to start a company in that field, I think I'd need to spend some time in the trenches of that particular field.

Likewise, I suspect if you do find yourself a lucrative and currently poorly served market niche to focus on, you will need to also find a few companies to work closely with, so you actually understand the problems and how they can be solved. Also, keep in mind that the harder existing software is to use, the more training existing users will have had to know how to use it. Don't underestimate how hard it is to convert someone from using "complicated but familiar" solutions to using "easy but new" solutions.

Not seeking to rain on the "I hate X" parade, just pointing out that if you aren't scratching your own itch, to some degree, you have a lot of work to do just to begin to understand the problem domain.

abalashov · on Aug 3, 2009

Thank you, very good post. Echoes my own thoughts.

It is a response that I have frequently entertained in my mind to entrepreneurial cheerleaders (and even sometimes PG's statements, insofar as they relate to solving truly niche commercial software problems, although I don't think he intentionally suggests that you can do it without the experience by any means).

But yes, I'm in telecom and I see outside consultants and vendors who are basically general IT in their orientation show up all the time and think that they can create better OSS/BSS (operational support/billing support systems) solutions, switch provisioning and monitoring software, and workflow/process management in the highly bureaucratised and rather regulated carrier space (i.e. Local Number Portability).

Then I see them get fired a few months later because they just haven't the foggiest earthly clue what minefield they stepped into. They would have benefitted from hiring at least one omnicurious person who has actually worked for a competitive carrier first.

jacquesm · on Aug 2, 2009

Linux. It's arguably one of the best unix clones ever and it is extremely stable. Still, I feel that another clone of macro kernel unix was a missed opportunity. Plan 9 and it's ilk (and QnX) seem much more the direction to go in for many reasons, most of them having to do with modularization, upgrading your operating system (even the kernel) without rebooting and user land device drivers.

Still, to make a living plan 9 and QnX are niches, plan 9 is very much an experimental os, and QnX is mostly relegated to being a real-time embedded systems os. If you've worked with either one of them for any length of time though you really will begin to appreciate that there are other possibilities that are worth researching.

Linus and Andrew Tanenbaum had this out a long time ago, and it seems that Linux 'won' by acclamation but I'm not so sure that wasn't simply pigheadedness instead of sound engineering reasons.

bayareaguy · on Aug 2, 2009

Some of my clients use Windows Server 2008 on remote colocated blade servers. It sucks because it's difficult to manage, it constantly raises useless authorization dialogs and it's terribly inefficient at even the simplest operations like copying files. To make things worse many operations have to be done using DRAC controllers and a broken IE interface over a slow network. Linux is a viable alternative and some are moving to it slowly but in many cases historic developers and programs are tied to Windows.

olihb · on Aug 3, 2009

I work in market research and have to use SPSS. It sucks because the old versions would crash for no good reason and the new versions are slow. The output is ugly as hell. I haven't switched to R, StatXP, Stata because I'm lazy, we have a huge number of old surveys in SPSS format and a whole lot of scripts and macro written.

Oh and the copy protection sucks. Seriously, if I want to pirate your software, I will download a crack.

dmix · on Aug 2, 2009

My mouse. I would love to control OSX entirely by keyboard. I tried the framing systems on linux but it just wasn't for me.

thunk · on Aug 2, 2009

Word. That itch inspired my current Big Scratch.

paulbaumgart · on Aug 4, 2009

Xilinx ISE. Admittedly I only used it for a college course ~1.5 years ago, and then only whatever version they made freely downloadable, but it was ridiculously buggy and user-unfriendly compared to any other development tool I've tried.