I find during the holiday season many pick up new books, learn a new language, or brush up on some other skill in general. Here’s my contribution to hopefully giving you a few new things to learn about Postgres and ideally utilize in the new year. It’s not in a top 10 list as much as 10 tips and tricks you should be aware of as when you need them they become incredibly handy. But, first a shameless plug if you find any of the following helpful, consider subscribing to Postgres weekly a weekly newsletter with interesting Postgres content.
The headline of Postgres 9.5 is undoubtedly: Insert… on conflict do nothing/update or more commonly known as Upsert or Merge. This removes one of the last remaining features which other databases had over Postgres. Sure we’ll take a look at it, but first let’s browse through some of the other features you can look forward to when Postgres 9.5 lands:
I recall extremely early stage where you’d build a feature, realize it was awesome, then the next day write a blog post for it. At some point you start to move from that to more coordinated launches. A larger coordinated launch allows you to reach a bigger audience, can lead to bigger deals, and help expand your overall market. But perhaps more importantly by the time you hit full launch you’ve message tested and ensured it’s going to resonate in the way you expect.
The process itself will both help amplify and validate/refine your message
This is often a more gradual process than a sudden single change, you’ll introduce new parts of this in time. And for many what an entire launch process looks like comes by trial an error, to help shorten that learning curve here’s key areas I pay attention for a launch and process followed by a rough timeline.
JSONB in Postgres is absolutely awesome, but it’s taken a little while for libraries to come around to make it as useful as would be ideal. For those not following along with Postgres lately, here’s the quick catchup for it as a NoSQL database.
- In Postgres 8.3 over 5 years ago Postgres received hstore a key/value store directly in Postgres. It’s big limitation was it was only for text
- In the years after it got GIN and GiST indexes to make queries over hstore extremely fast indexing the entire collection
- In Postgres 9.2 we got JSON… sort of. Really this way only text validation, but allowed us to create some functional indexes which were still nice.
- In Postgres 9.4 we got JSONB - the B stands for Better according to @leinweber. Essentially this is a full binary JSON on disk, which can perform as fast as other NoSQL databases using JSON.
First some background–I’ve always had a bit of a love hate relationship with ORMs. ORMs are great for basic crud applications, which inevitably happens at some point for an app. The main two problems I have with ORMs is:
- They treat all databases as equal (yes, this is a little overgeneralized but typically true). They claim to do this for database portability, but in reality an app still can’t just up and move from one to another.
- They don’t handle complex queries well at all. As someone that sees SQL as a very powerful language, taking away all the power just leaves me with pain.
Of course these aren’t the only issues with them, just the two ones I personally run into over and over.
In some playing with Node I was optimistic to explore Massive.JS as it seems to buck the trend of just imitating all other ORMs. My initial results–it makes me want to do more with Node just for this library. After all the power of a language is the ecosystem of libraries around it, not just the core language. So let’s take a quick tour through with a few highlights of what makes it really great.
These days if you’re creating a company you likely hope to accomplish more with less people, two ways of doing this fall to: The sharing economy and creating a platform. It’s easy to see the case for this when you have such unicorns like AirBnB or Uber. The opportunity for each of those to compete against hotel chains or taxi services which each need to manage their own inventory is incredibly exciting and revolutionary. In a similar fashion platforms can offer much the same, Heroku’s platform and marketplace made it easier than ever for developers to click a button and get everything they needed years ago. It’s not just their code, it’s everything from Postgres to Mongo to Logging. Or take the app store as example. Smart phones weren’t a new thing when the iPhone came out, but it was only the saviest of users that had apps installed on their windows smartphone or blackberry. The app store made the iPhone different than any other phone by allowing others to build and improve it, turning the iPhone not into a phone but a platform.
When it comes to go to market and marketing there’s lots of pieces in a toolchest that all work together. One that comes a bit later, but if used properly (much like a PR agency) can be valuable is industry analysts. And while working with a PR agency can quickly start to become clear. How to work with analysts so it is productive on both sides can take a bit longer to figure out, or at least it did for me. Even before you do start working with them there’s the question of if or when should you. Here’s hoping this primer makes it a bit faster and easier for others.
You’ve built your product and you’re now ready for your first major launch. Or you’ve been through a launch or two, but are looking to scale the process as you’re doing more launches and announcements. You really have two options: do it all on your own, or work with a PR agency. One frequent crossroad is that you’re not at the point of a full time PR person, but unsure what a PR agency can offer you; and, further what’s the best way to work with them so you’re getting the maximum value.
As I’ve talked to more startups lately, it’s become clear that effectively working with PR teams and the media is mostly learned by doing. Because there’s not much guidance out there, here’s an attempt at some basic guidelines.
Often when you’re tracking a metric for the first time you take a look at your average. For example what is your ARPU - Average Revenue Per User. In theory this tells you if you can acquire new user how much you’ll make off that user. Or maybe what’s your average life time value of a customer. Yet, many that are more familiar looking and extracting meaning from data median or a few different looks at percentiles can be much more meaningful.