HNHacker News
TopNewBestAskShowJobs

greenyouse

258 karma · joined August 21, 2012

ed@edbabcock.com
submissionscomments
greenyouse··on Visual Studio IntelliCode
I remember reading when Facebook announced Aroma to auto suggest common coding patterns to FB engineers[0]. I haven't used much for code completion before but this idea sounds really interesting. It sounds like the IntelliCode implementation sources suggestions from repos with a higher number of stars so hopefully that helps with noise. For JS I'm sure the variability in style is pretty big.

For people that have been trying IntelliCode for a while does it help much with development? Does it have the ability to suggest larger code snippets or only word by word?

[0]https://ai.facebook.com/blog/aroma-ml-for-code-recommendatio...

greenyouse··on Ask HN: Gmail / G-Suite Alternative
You could also DIY with a VPS and an email project like iRedMail, Soverign, or others. I used iRedMail for a few years and it was pretty easy. As a potential benefit, if you want to pick up more DNS record + email knowledge and have a go at maintaining a linux box it's not a bad project. It's ~$2/month depending on what you use.
greenyouse··on Host Your Own Blog with Gitlab and Netlify
I guess this is might be easier to set up but you could get something similar on AWS by posting to an S3 bucket and putting pages into CloudFront CDN. Adding CodeCommit could automate deployment for $1/month or you could write a script to publish new pages for free.
greenyouse··on Ask HN: What overlooked class of tools should a self-taught programmer look into
This might not be for a hobbyist since the setup is difficult but validating your frontend work with automation is very interesting. I didn't realize for a long time that I could use visual regression tools and Selenium at the end of development to catch bugs. I'm surprised that there aren't more tools in this space targeted for developers. Selenium Grid with WebdriverIO is incredibly helpful. If you put together a visual regression system for validating your daily work you could save some serious time.
greenyouse··on A World Run with Code
Just for fun here's a link to a 2010 flash crash before those rules were in place where a trillion dollars were dropped in half an hour due to algorithms.

https://en.wikipedia.org/wiki/2010_Flash_Crash

greenyouse··on Ask HN: What’s an app that you wish you had?
There should be a documentation system that suits the role of each job role, group, and company division. There should be something better than the landfill of data that are wikis or confluence.

Why can't there be a system that summarizes the details of apps based on README files or architecture files? Supporting documentation should be personalized to the user so if you're a developer you'll get API info and sequence diagrams (product managers could get product notes, etc.).

The documentation system should be alive with interactive examples that drive home what the experience is (embedded invision, web pages, swagger docs, etc.).

Documentation should be automatically deprecated as the state of your system and product change. Archives of old stuff could be kept for reference but the actual interactive system should be only live things so they're relevant.

Does this already exist?

greenyouse··on Web Components could replace frontend frameworks?
Browser specifications are not nearly as transient as proprietary frameworks. Web components will be here next year, five years from now, ten, and thirty years from now. We should focus on how we can take advantage of the new spec to enhance existing frameworks.

Popular projects like React use templates but require you to bundle the entire framework and the speed of the vDOM diffing happens at the speed of JS. What if it could take advantage of HTML templates to work at the speed of C++ by leveraging built-in browser support?

One of my hopes is that we can stop the bad trend of rewriting the same front-end components over and over again. Instead there could be a set of canonical components that people use for different web site types. If you're building an e-commerce checkout system, you could pull in the email address input built by Twitter, Stripe's credit card input, and the address input built by some person on GitHub that has 5,000 stars. Authors could leverage CSS properties and ShadowDOM for nice style hooks so developers could customize the branding.

The job of web frameworks could shift from "how do people build UI components" to "how do people build SPAs with routing, cookie auth, service worker caching, etc.". Not that current frameworks don't try to handle app level things but there is a lot more that could be done to make this easy for developers.

It would be nice to see web components as orthogonal to framework UI styling. We should be able to pull in web components from anywhere and frameworks should be smart enough to co-opt the components by treating them as native DOM elements (like div or span).

Sorry if this was sort of a rant I didn't have much time before work...

greenyouse··on Ask HN: What music do you code to?
SomaFM groove salad is great! I feel like since I only listen to it while doing work it snaps me into work mode. It's easy to zero in and forget that the music is playing.
greenyouse··on Sensible Software Engineering
You can tune the pixel difference in Selenium testing so it doesn't fail when under the acceptable pixel change limit. If it's off by one or two no biggie but off by 30 or 50 then maybe there was a structural change to the page, so it should fail. Something like "misMatchTolerance" from wdio-visual-regression-service[0] would allow for tweaking this. I'm sure there are similar tools for other languages.

If the UI tests catch bugs during development or help the team during a data migration, they're probably still worth having.

[0] https://github.com/zinserjan/wdio-visual-regression-service

greenyouse··on Ask HN: What are some good back end website ideas?
These might be too difficult but just for fun here are some ideas for mostly backend.

Personal Music Tracker - You can get new music ideas by scraping content from a community radio station playlist like KCMP (https://www.thecurrent.org/playlist). Run it against a personal database of songs you already know and have the system send you a daily/weekly/monthly list of songs you haven't heard. It's fun to build.

Phone Proxy - Set up a disposable number for yourself on Twilio that will proxy incoming calls against a whitelist of approved numbers. You could do Twilio Studio to build an IVR that prompts the caller to press one to route the call. If they stay on the line you could do fun stuff like transfer to an It's Lenny server or something. It's not difficult to program and will teach you about some phone technology.

IoT Blog Reader - Follow blogs that you like and have them read to you via common IoT devices like Alexa/Google Home. Like an RSS reader but built for speech. Use an RSS reader to pull articles, run transcriptions of the articles with trained ML TTS models, and store the data with SQL. De-dupe the transcriptions based on URL. Use pre configured ML models for text to speech like AWS Polly or GCP Wavenet. Probably way too hard...

TextToSpeech for books - Upload a book. Get emailed the audio transcription. Use pre trained ML models for TTS conversion. Maybe use a SASS company like mailgun for the email part so you can just focus on extracting text from different file formats? Could start with plain text files initially.

Story Point Estimation - Use ML (or basic statistics) to estimate how long JIRA tasks or projects will take to complete. Use a general corpus of data to train an ML model, then feed it a team's data so it can adapt accordingly. There are some whitepapers on it like https://arxiv.org/pdf/1609.00489.pdf but you could look around for easier statistics algorithms to use instead so it's easier to build.

greenyouse··on Ask HN: How Can I Be More Hardworking?
I think the most important thing is wanting to do something. The problem in front of you shouldn't just be work - you should actually be interested in solving it.

That would probably range from something that's so easy you just want to drop the hammer and knock it out to a massive problem that you'll spend months or years chipping away at. Either way if you really are interested in working on it you'll think about it in the background and approach solutions from multiple angles. When you're excited about working on the problem you'll have no issues with working more than 10 hours per week if that's what it takes.

Everybody is different but I've found that when a problem is actually fun to work on "productivity" techniques don't really do much. It's mostly making a plan for what to do, working a bit, learning more, and either solving or starting the work cycle over again with new information.

So I think you probably need to work on more problems that you're actually interested in.

greenyouse··on A small notebook for a system administrator
I'd vote for getting a disposable laptop with good enough hardware to do most sysadmin things. A 2006/7/8 macbook isn't a bad choice IMO since it has all the standard ports. It supports two disk drives and the keyboard works OK.

If you poke around you can usually get one in the $20-40 range. No USB-C or large processing power, but if you're mostly networking into other computers it shouldn't really be a bottleneck, right?

greenyouse··on Ask HN: How do you organize everything you want to do?
If you're lacking time to work on stuff you could keep track of your daily tasks and see trends in how you spend your time. If you know what you're doing, then you can re-organize your schedule to do things that are more important for you.

If you have a large list of projects to do, you could try assigning value to each and put them into a queue.

Take a spreadsheet, put the projects into rows, add columns for what's important to you, weight the columns appropriately, compute a final importance score, and rank the items in the queue. Column names for projects could be: community value, work value, personal value, and maintenance cost (negative). That at least datafies the tasks so you can objectively work with them.

Otherwise org-mode for taking notes and small daily tasks. Having random facts on hand about systems at work and what you did 5 months ago to solve problem X have been really helpful for me.

There is something to be said for just getting excited about something and hacking it though. It's way more fun :D

greenyouse··on Ask HN: Is there any work on self-testing software?
I don't think there's really a thing like you want where the system changes it's parsing logic on the fly. You could definitely set up acceptance tests for the system to see whether your program can parse the Nevada legislature data properly as it changes. The program could pull the data fresh from the API, run through, and check that the end result is the same. Automation tests can run on a scheduler so you could at least figure out when the data changes (daily, weekly, etc.).

Could you post back if you do get the dynamic parsing to work in Haskell?

greenyouse··on Ask HN: Can you help me find a previous HN post?
https://hn.algolia.com/ is nice for finding old posts too
greenyouse··on Little Languages
I worked through the CTM book[0] a while ago and I think this was the general approach taken using the Oz programming language. Each programming paradigm was split into a kernel language that would demonstrate specific programming features. I'd highly recommend trying it out.

I really like the diagram they use to split up the languages + paradigms[1]

[0]https://en.wikipedia.org/wiki/Concepts,_Techniques,_and_Mode...

[1]https://www.info.ucl.ac.be/~pvr/paradigmsDIAGRAMeng108.jpg

greenyouse··on Slow Software
The only nitpick I have is that the sidebar comments are built in a weird way. They're nested in the middle of the paragraphs inside small tags. It would be a little nicer if they split the sidebar text into aside tags that come after the paragraph.

It's refreshing to see a site that loads quickly though. That would've been a bit hypocritical otherwise :p

greenyouse··on Ask HN: Why don't people program on their phone?
You could build a slide out keyboard for yourself and attach it via the usb port on your phone or via bluetooth. I've seen a wireless keyboard attachment for iPhones already. Maybe to get a larger keyboard you could have a well designed fold up or rollout keyboard that is compact?

You might also be able to use your phone as a thin client if you can put up with the limited screen space. There are some good systems like Apache Guacamole. It's too hard for me to use but maybe it's possible with a phablet or diy smart glasses and a bunch of determination. :)

greenyouse··on Kubernetes for personal projects? No thanks
If you go to joyent.com you'll see "Multi-Cloud Kubernetes" is on their homepage above the fold and listed in their products & services category. SmartOS is an operating system. Triton is a data center for running container instances. Kubernetes is a container orchestration tool. These are all solving separate problems...
greenyouse··on Ask HN: How do you market yourself at your mid-30's?
You could try quantifying the number of jobs available for different technologies. Then match up one of your existing skills with a large portion of the market (I would guess PHP + Python would be easiest because the others are either niche or have more difficult hiring processes). Settle on getting one particular type of job (e.g. Magento PHP developer), study a bit, and try practice interviews online while you apply.

Worst case scenario the companies won't like you and you'll be turned down.

It sounds like the hiring managers are getting mixed up by your resume. If you aren't already doing this, you should focus your resume for the specific job and leave out years of experience in other non-applicable technologies. It wouldn't hurt to include a few links to your open source work in the resume if you've been coding in X language for years (you could try contributing to a popular open source project too but that might be tough). There are likely resume review processes you could use that could give you advice (btw, feel free to post a link if you already have it in digital format and I'd check it out).

For getting in the door with a junior/mid level job you should try using a recruiter. They get a bad rap but their interests mostly align with yours at this point. They get paid when you land a job so they'll hound employers pretty aggressively and fight to bump up your hourly rate. Try updating your LinkedIn profile with a format tailored to the type of job you want to get and they'll probably reach out to you. Already having 2 years of development experience is good.

People have varying opinions on whether to multi-task during your job search so you could hold down a side job or do freelance work. Since you haven't been in the industry for a few years I would probably suggest just going all in on prepping. Focus on what technology you want to hone, tailor your profiles for that, practice interviewing/project building with that, and put applications out.

greenyouse··on Reading with a pencil
Sure, I usually set org files up by book/talk > chapter/concept > (any further level of nesting)+. Here are a couple examples. From watching "Abstract" on Netflix I did one entry for the show with a bunch of random notes because there wasn't as much structure to the content. Most of my book notes more or less follow the book > chapter > sub heading pattern. Tech talks usually have a format like talk title > subject > details. I've stuck to using one file for most of my notes the last few years. You should play around with it and see what works best for you.

Reading with a desktop or laptop has been easier for me since it's more familiar. I'm not really a fan of mobile phones or kindle readers. Evince or fbreader work OK on linux. You could add espeak or another TTS tool for speaking text. On OSX there's acrobat with voice over. Screen readers exist but are harder to control. The tools for reading via TTS are pretty primitive right now unfortunately.

I'm pretty bad at syncing notes... I've primarily done it with rsync over ssh. If you save things back to a server and sync from there that could work. Otherwise there's easier stuff like Google Drive or Dropbox.

greenyouse··on Reading with a pencil
Do others do a similar thing but with digital notes?

I used to take notes on books in notebooks and mark the page numbers. Over time it switched to ebooks with TextEdit, then org-mode for writing thoughts on books/talks. It's nice to break a book down by section or concepts and have it all in one place as opposed to being scattered throughout random pages or listed at the end of the book.

It's usually pretty easy to convert ebooks/pdfs to audiobooks or use a TTS tool to read each page, one at a time. That makes it easier to write notes while you're listening to the book. You can also get used to reading around 400 wpm, which is nice if you're a slow visual reader.

Since the comments aren't directly in the book, it's not as geared towards English editing as marginalia is. It's great for summarizing technical material though or learning new concepts!

greenyouse··on Ask HN: What is the best API documentation you have ever seen?
Stripe is probably the best but I think Twilio is a close second. For both the page designs are easy to read, there are clear examples (some are even interactive), and the sites are easy to navigate.
greenyouse··on Designing a Good DSL
Would anybody mind sharing links to good DSLs?

From Clojure I think hiccup is a good example.

https://github.com/weavejester/hiccup

greenyouse··on Lessons learned scraping 100B product pages
I'd agree. As somebody scraping content, what's so bad about increasing the timeout to like 10 seconds? That way the servers can handle the traffic easily and you're not being a jerk. If you have one async thread for each domain you can still get lots of data quickly. Causing a denial of service attack is very avoidable.
greenyouse··on Something about functions in Bash (2017)
On the data processing with bash part, Joyent Manta always looked interesting[0]. Reading about data processing via bash makes it sound pretty fun.

[0]https://github.com/joyent/manta/blob/master/README.md

greenyouse··on Ask HN: What are the things that you have automated in your personal life?
There wasn't much to the paper archiving. I bought a fancy scanner[0] and already had a shredder.

To store the files on disk I keep directories marked by topic/year/month/day. The topics are stuff like taxes, finances, bills, etc. The important documents are in a couple encrypted HDDs that I use for backups. I've been thinking of doing an offsite storage like Backblaze B2 but it's probably overkill.

After scanning stuff in, everything gets shredded. There usually aren't many things to save but once they've been scanned they're there forever.

[0] https://www.amazon.com/Fujitsu-ScanSnap-iX500-Duplex-Scanner...

greenyouse··on Ask HN: What are the things that you have automated in your personal life?
Digital archiving of important physical mail is a nice way to go paperless and keep track of documents. Get a good scanner and shredder, then set up a system to track your scanned documents. No more misplacing files and having to stuff cabinets with papers.

Proxy phone number via Twilio or Google Voice for the times when you'd rather not give your real number out.

Audio translations for digital books using espeak + pdftotext + shell scripting. If you didn't know, espeak can output wav files. It's a nice way to read since my commute is 3 hours a day.

Recipe management for putting meals into rotation and generating a shopping list. Check out Gourmet or Krecipes on linux. It could probably be hooked into a grocery delivery service like Prime Pantry or just migrated to one.

General automation with Twilio + serverless cloud functions is pretty good. There's a new Twilio tool for building IVRs (interactive voice response trees, like those phone menus you get when calling a cable company). The super cheap on-demand pricing for cloud functions makes this basically free to run.

I started a personal SMS/Voice service for some of my tasks like what's on [favorite radio station], Bus line directions, shops near me, etc. There's no real point yet but I feel like it could have uses that are actually helpful...

greenyouse··on Ask HN: What's your favorite way of getting a web app up quickly in 2018?
Serverless + managed cloud services for heavy lifting (cognito, dynamodb, cloudsearch) + basic logging, alerts, tracing stuff

static site or React on frontend

Pretty quick to roll out and free to run if the app never takes off. If you need to scale later, you can build your own serverless versions of some cloud services to save money.

greenyouse··on Ask HN: What are your criteria for selecting cofounders?
I really liked this blog post from Leo Polovets for measuring the overall risk of a startup.[0]

Maybe in the same vein you could do it for cofounders? De-risk yourself by picking a cofounder that you know well and will be helpful. It's definitely not perfect but maybe this will help...

If you pick someone that you know well, you'll be a better judge of whether you'll have work conflicts later. I'm not a fan of working on a project together to test your chemistry. It's still a big gamble if you do that. People are too nice.

Your cofounder judgment ability:

- random person

- acquaintance

- previous co-worker

- friend

- family

You could draw up the major roles that you'll need to take on for the early stage part of the startup (UX, sales, engineering, design, marketing, management, fund raising, etc.). When picking a partner try to optimize for filling as many of the roles as possible. It helps if they have a different background but I think proven skills are more important.

Cofounder skill balance:

- they're me and we do all the same stuff

- same skills but different background

- they know some other skills but together there are still many gaps

- together you cover most skill gaps and have different backgrounds

- they know enough of what you don't that together you'll cover all the skill gaps

From the roles you laid out, how much experience do they have with each? Add up the points 1-5 for each job to get an idea of how likely they are to be good at their tasks.

Cofounder job experience:

- never done work like this before

- worked a bit in the industry but not this role

- worked in this role in a different industry

- worked in this role in this industry

- excelled previously working in this industry doing this job (really know the ropes)

Take these three criteria and weight them properly for a clunky formula that sorta helps with cofounder picking, e.g. 5 * chemistry + 3 * skills + (total_experience / # jobs)

non-starters:

- can't do it full time

- squabble over equity or other founding stuff

- don't want to do work (after you agree on tasks)

- talk past each other often/can't work together

I'd second what other people in this thread have said about reading The Founder's Dilemma. It's well worth the time if you haven't read it.

[0] https://codingvc.com/how-to-de-risk-a-startup/

← PreviousPage 3 of 5Next →