3 ms·
I know this is a tough sell but I am going to say it anyway. Automation effects every single person and every single job. For example, if your job is writing
by ilaksh 12y ago
I know this is a tough sell but I am going to say it anyway. Automation effects every single person and every single job.
For example, if your job is writing custom software, it becomes progressively harder to find problems that don't already have open source solutions.
Or look at how component marketplaces and open source components or modules increases software reuse. There is not an infinite capacity for software complexity to increase. So that reduces the amount of work to be done simply because solutions are being shared rather reinvented.
- diltonm 12y ago>> For example, if your job is writing custom software, it becomes progressively harder to find problems that don't already have open source solutions. Open Source is just software, solutions take people. It will continue to be the case for the next 50 to 100 years. The computer industry is still a baby learning to crawl.
- bjelkeman-again 12y agoAt the same time that means we can work at a higher level of abstraction. I see this daily in my work. There is a load of work to be done in providing tools for the none-Western world and that work has barely started, for example. Or environmental management, which hardly uses modern IT tools yet. I am not afraid of running out of work in my lifetime.
- tluyben2 12y agoCurrently this is only working to some extent; there is no sign of software development demand going down because of open source or saturation. It's going up, and fast. I would guess the reason for that is that we don't have any way of truly composing those open (or closed; it doesn't matter for your example; closed source works in the same way; if you could buy it you would) source solutions to do something a company needs. You have to find them, test them, hook them together, test them a lot more, deploy them, test them a lot more and then change them over time. Most companies fail at finding them (and write them themselves); if they do, the first testing is done badly, so after hooking them up the solution won't work as expected etc etc. There simply is no robust way of composing 'software' easily without actually blowing your budget (a few times over) to see if the solution(s) fits. Intuitively I can see that, feature wise, SugarCRM, Solr, Hadoop, HBase, Postres and OpenERP would be a fit for your company if I could integrate them in a certain way; Google gives you enough clues for people who did this (years ago), so how hard can it be? Hard enough to spend massive amounts of money and then still not even be slightly happy about it. Happy meaning that your 1000s of employees can use it and your company runs on it as you, as CEO, envisioned. Is it more or less costly than writing only what you actually need from scratch? No idea, no-one is going to try that; you try SAP/Oracle/MS or this kind of thing and that's it.
- brandonmenc 12y ago> For example, if your job is writing custom software, it becomes progressively harder to find problems that don't already have open source solutions. There's a lot of work to be done writing glue code.
- CmonDev 12y agoThere is a whole business of 'integrating' the software products that are kind of ready (much more so than OSS) - ERP (Dynamics, SAP etc.).
- Padding 12y agoFor better or worse, this isn't true. Software development to some extent is a creative process, thus eluding automation in many ways, though not for lack of trying. Much like book authors, in the general case, can't really "re-use" subplots of one another's stories, software devs can't really, in the general case, re-use one anoher's implementations. Of course there are many more specific situations in which software may be re-usable than there are for subplots in books .. but that doesn't impact the overall argument. Much like a detective story will have some crime being commited by some perpertrator, harming some victim and being investigated by some detective, an e-mail client will run on some OS have some GUI and exchange some IMAP and SMTP messages. But the creation of the specific detective story and the specific e-mail client still require creative input by some human before they can come into existence. Or is your argument that there is less need for e-mail clients and detective stories because there already are enough of them in existence? If so, then this is an issue orthogonal to automation itself.