5 ms·
Ask HN: How FULL is the stack?
I consider myself to be a full stack developer in that I can write backend code (node.js,Java Spring, PHP) and frontend (Backbone.js, Ractive.js). I am not really good with CSS though.
In the future, do you think a full stack developer would be expected to be familiar with machine learning too in Python or something else?
- stephenr 11y agoin the future, a full stack engineer will be required to do everything. - Invent/discover new materials to replace silicon, copper, gold, etc - Create CPU using above materials - Create Kernel to run on above CPU - Create Software language(s) for application - Create Software libraries for above language(s) - Create application using above libraries and laguage(s) - Design interface - Create new interface rendering language - Create new image format - Create above interface using above rendering language and image format - Create new network transfer technology - Create new Ultra Quad Super Duper HiDPI physical screens to take advantage of new interface - Create new GPU to drive said screen - Create new batteries to power above devices - Create new OS to use and manage above devices - Create new humans with eyeballs capable of appreciating Ultra Quad Super Duper HiDPI screens - Create pancake stack for above humans to eat for breakfast. Seriously though - if a project has the same person writing Database queries and Media queries, run. Run like the wind and do not look back.
- arisAlexis 11y agoyour about reads "Ops/infrastructure and web app development" :D
- stephenr 11y agoYes, it does. Also, I was once employed primarily to write HTML/CSS/JS.
- dudul 11y ago"Full stack engineer" is the new "Coding Ninja". It doesn't mean anything. It is most often used by non-technical founders who just want to hire one guy to do the work of 3. As you said "How full is the stack?" Do you need to know how to setup and administer servers? Do you need DevOps skills? Do you need to be able to write backend/frontend code? Do you need to know how to design the UX as well? I systematically ignore all positions for "full stack engineer" because in my mind it translates to "we don't know computers but we have a Great Idea(tm) and we just want some nerdy guy to handle all these computer thingies for us".
- dozzie 11y ago> "Full stack engineer" is the new "Coding Ninja". It doesn't mean anything. > Do you need DevOps skills? Certainly, as the "DevOps skills" means the same as "full stack engineer": nothing.
- dudul 11y agoWell, DevOps is definitely an over-used buzzword, but I think it's a good term to use as an umbrella for technologies like Chef, Puppet, Ansible, etc. It is definitely a different skillset than the one required to write software. I think DevOps gets a bad reputation mostly because of companies who just use it to look cool without really embracing the "infrastructure-as-code" philosophy (a bit like Agile/Scrum "we do Scrum, but...")
- dozzie 11y ago> I think it's a good term to use as an umbrella for technologies like Chef, Puppet, Ansible, etc. No. Those already have their umbrella term: infrastructure as a code. You even mention it yourself.
- davelnewton 11y agoThat clearly takes too long to say. DevOps covers more than just "infrastructure as code", though.
- PaulHoule 11y agoI would say it is a viewpoint I attained from doing web projects for various "startups" and "internal startups" starting in 1999 or so. In these cases you have very small teams so you can't afford to have separate DBA, Sysadmin, Network Operations, and Development people -- at least not entirely. When you've had experience doing everything yourself I think that makes you a better team member. For instance I write very solid server code because I hate being woken up at 3AM.
- PaulHoule 11y agoMachine Learning, as it is practiced today, is a field with a huge amount to know -- just as you could make webapps with plain HTML or write them in C and compile them to Javascript you could choose a huge number of algorithms and specific implementations of algorithms. A big part of the future is systems like this https://aws.amazon.com/machine-learning/ https://aws.amazon.com/machine-learning/ which let you get machine learning done without even choosing an "algorithm" such as neural network, decision trees, support vector machine, whatever. The thing either has a very flexible algorithm in it, or it chooses an algorithm. That kind of thing is an "expert system" which is an expert on machine learning -- it tries a bunch of things, evaluates them, and shows you the best one. A better tool would engage you in a dialog over the data and not be such a black box.