4 ms·
Considering the quality of the code I see currently and the costs created by that technical debt I'm not sure if becoming less elitist will be a good thing. Ins
by flukus 10y ago
Considering the quality of the code I see currently and the costs created by that technical debt I'm not sure if becoming less elitist will be a good thing. Instead of all these code camps, I can't help but wonder if we could get 90% of the value and 10% of the problems if we taught people shell scripting instead of programming.
It looks like wireds ad-blocker blocker has been circumvented now at least.
- vinceguidry 10y agoThe nice thing about software projects is that they're easily replaceable. They can be treated as any other business asset in that you can hire a shop to develop them or do it in-house. So long as you have enough business eyes on the process then it works. Expensive, but it works. Businesses are used to expensive. The real value of good code is way, way less than developers would like to believe. This is coming from someone who is damn near a perfectionist about code. I have to constantly check myself and ask whether whatever I'm demanding we spend time and effort on is really worth it from a business perspective. Occasionally it is, most of the time it's not. If coding wasn't the thing I could make the most money with at this point in time, I'd do something else and code free software as a hobby. At least free software really benefits from great code practice.
- the_cat_kittles 10y agoas a general rule, i think the quality of the code matters more as more things depend on it, so i think you need to temper your statement a bit- the value of good code depends greatly on what its being used for.
- flukus 10y agoI've never seen management buy in for a rewrite, especially when the code is so bad that the company starts circling the drain. The only time I've seen it is when there's been a fundamental technology shift, like from VB6/C++ to java/.net. Network effects also make it hard. Sometimes you can't upgrade X because A, B and C feed into it and D, E and F feed out of it and they're all tightly coupled (another thing that's even worse with less skilled programmers). Replacing a single component can mean rewriting the software stack of the whole company. > If coding wasn't the thing I could make the most money with at this point in time, I'd do something else and code free software as a hobby. At least free software really benefits from great code practice. Completely agree here. The money is the only thing keeping me in the industry, if that went away I'd do something much more blue collar.
- vinceguidry 10y agoSometimes I wonder if the money is the only thing that keeps me coding. It seems too hard to do well otherwise.
- braveo 10y ago> The real value of good code is way, way less than developers would like to believe Soooo much this. But it's also important to note the difference between code and system design. The vast majority of code that's written is solving problems where system design doesn't tend to make a huge amount of difference. It's only when you get to a certain scale and/or in certain domains that the system design starts to matter, and even there, the code itself is of little import.
- collyw 10y agoThen with shit code and even more for shit system design you just spend far longer maintaining the technical debt rather than adding genuine value. Every feature takes 4 times longer to implement because of it. It costs more, and ships later. And there are always bugs to fix. Its obvious to anyone capable of writing decent code that it does have value.
- vinceguidry 10y ago> Its obvious to anyone capable of writing decent code that it does have value. Business users are the ones that have to make the value decisions, and they can't write any code, much less decent code.
- flukus 10y agoAnd they are oblivious to the costs incurred by shit code.
- braveo 10y ago"shit code" is in the vast minority of code written, most of it ranges from merely crappy to so beautiful you'd auction off your own mother just to bask in its presence a little longer The reason architecture starts mattering is because you can do things like put that "shit code" behind an interface/microservice/whathaveyou and then the rest of the system just doesn't care. This is why code itself isn't particularly valuable in the way you seem to be implying. It isn't without value, but while software is valuable, it's rare for specific code to be where that value is at.
- jasonmp85 10y agoAlso: shell scripting is programming.
- dsacco 10y agoI think the article is trying to argue that we should embrace the bifurcation of the tech industry. It's pretty fair to say most software developers don't need a full computer science education. In my view, web development is especially vulnerable to continued commoditization. It makes sense to optimize the lower hanging fruit among programming jobs - aside from perpetuating high (arguably, inefficiently so) salaries, the barrier to entry really doesn't need to be as high as it is now. That's not to say coding bootcamps are the answer, but vocational training is at least a step in the right direction. From one perspective, this would be good because driving down barriers to entry will theoretically increase jobs and reduce costs. From another, it's obviously not great for web developers who are highly paid, though it could theoretically raise salaries even higher for the fewer highly trained engineers who actually require a computer science education. Either way, I think the industry would need to substantially change before the less complex/more rote programming jobs could be subsumed into vocational training. It would require a pretty significant cultural shift.
- qq66 10y agoBeing less "elitist" does not mean that standards are any lower. For example, electrician is not an elite or elitist profession, but the quality of electrical work is very high, because there are strict standards, developed carefully, that all electricians adhere to. It is because electricians do not improvise on the job that their work output is high quality. Experts develop a standard toolkit to deal with 99% of the electrical work out there, and electricians deploy those toolkits with a much greater focus on standardization, conscientiousness, and correctness rather than creativity or novelty.
- sandworm101 10y ago>> shell scripting instead of programming. I was at a discovery conference (lawyers) listening to a vendor pitching the cost of writing some ediscovery syntax to delve through a 50gb database of emails. While he was talking I performed a grep search on my ubuntu netbook, with the database on an encrypted thumbdrive. In a matter of minutes I convinced the client that he didn't need coders to do much of anything, that an idiot lawyer with a laptop could do it from his hotel room that night (which I then did). The vendor left the room very angry with me. The downside was that rather than the 100k I saved the client, all I could ethically charge was my rate for 6hours of overtime.
- flukus 10y agoI always love these stories. I'm pretty sure half of the over engineered apps where I'm working could be replaced by a handful of shell scripts. One particular one recently had us adding a whole lot of tables to the database with the accompanying sql and code so we could monitor performance in real time. The information could have easily been generated from our log files with a sed script or a flex program.
- sandworm101 10y ago(without breaking at-client priviledge) The full task wasn't exactly simple. It started with a failing part in a product. The issue was when exactly the manufacturer should have been aware of the issue. All we had was a database of emails to and from local reps, who as it turns out had been fixing the problem for a long time by telling customers to disable certain features. This fix caused other damage. From the email pile I created a graph showing when customers emails started to mention problems, when the first reps started telling them of the fix, and when those reps started talking to each other about this fix. My client-manufacturer then had an idea of when his organization should have gotten the message and put a stop to things. This then informed the negotiations. So it was cross-referencing a list of vendor emails addresses, customer email addresses, some keywords, and putting it into a timeline. Bash bash bash, a solid night's work but nothing needing actual code. Had this gone to trial we would have had the ediscovery vendor come in and do a full and proper report. It didn't, and both sides looked took my report as useful enough to come to an agreement. (I gave them my bash scripts and they verified my conclusions themselves.)