Dave’s Tech Ramblings

How to Prototype in a Design Sprint - Meetup

2018-03-20T03:59:49+00:00

How to Prototype in a Design Sprint

This is the meetup I attended and I wanted to get out a few of the key points and my take on them down.

This is the synopsys:

In a Design Sprint, prototyping day is Thursday. This is where you will create a testable prototype for Friday’s user tests.

Thursday is about illusion. You’ve got an idea for a great solution. Instead of taking weeks, months or, heck, even years building that solution, you’re going to fake it. In one day, you’ll make a prototype that appears real. And on Friday your customers will forget their surroundings and just react!

Prototypes are easier to build then you think. These days, thanks to the internet, there are a crazy amount of easy to use prototyping tools available. Most of them free! In this session, we will show you some of the tools that we use when prototyping and share our experiences of prototyping best practices.

The prototype mindset:

You can prototype anything

Prototypes are disposable

Build just enough to learn, but not more

The prototype must appear real

I’ve already listened to the audiobook, “Sprint” by John Zeratsky, Jake Knapp, Braden Kowitz, and that is what this meetup was about.

If you don’t already have an Audible account you can have the book for free:

It’s totally free and you won’t need a credit card if it is your first time accepting an Audible book from a friend.

Get it here

After you accept the book, you will be prompted to download the Audible app to start listening.

The first speaker Rohan Perera @Rohan_LD, one of the organisers, gave an introduction and his main points were as follows.

At the Start

Start with the end end mind.
- What are you trying to test with this sprint?
Finding out you should go further is a big win
- It is much “cheaper” than building a full solution before you find out it’s not going to work for your users
This is about creating a tangible discussion with your testers
Getting started is better than being right in the beginning

What is prototyping about?

Think of it as a façade of the product (or a portion of the product). It is an impression of what a future reality could be like. It’s all about learning, good and bad feedback are the goal.

Prototyping Options

Paper (including Keynote, PowerPoint, Word, PDF, etc.)
Service (as in changing an organisation)
Physical Space (changing how people interact with and office/shop/etc.)
Object (using 3D printer, flyers, leaflets, etc.)
Explainer Video Some examples given
- Dollar Shave Club
- Dropbox (both their original video and the latter video that propelled them to a billion dollar company)
Website
Tablet/Mobile App

The second speaker Daniel Alb @danielalbro gave a talk entitled “10 Golden Rules for Efficient Prototyping”

1. Make it testable

If you can’t test it, there isn’t any real value. The testers need to touch it and feel like it’s a real world situation.

2. Make it believable

It should feel like the real thing
You need it to have a beginning, a middle, and an end
- How do they start, what is the path, and where do we have them stop
It need consistency in content
- don’t show shoes in the shop screen and dresses in the checkout screen
Don’t break the illusion that this isn’t the real world
- Use real content and real images not place holder details

3. Words are worth a thousand images.

Write your copy first, it matters very much. Then you can make the images fit with the rest of the prototype.

4. Focus on the flows and scenarios

The flow and scenario is more important than making it look as pretty as possible
- User personas will help with this
- Remember every call to action is a different journey you have to build and discuss with the tester

5. Prototype only what you need

Only do what you need to do to answer the question(s) that are the focus of the sprint

6. There is no high fidelity or low fidelity only the right fidelity.

Make it as nice as it needs to be. A sprint is a fixed amount of time, so some things will have to be a trade off.

🗝 7. Don't get emotionally attached to the prototypes

These are tools for testing like a scrape of paper to guide you some elements will be kept some will be changed and some will be tossed out.

8. Kill ambiguity and get on the same page

Documents can be misinterpreted, but experiences are shared (for the team and the testers)

9. Fail cheaply

One of the main goals it to avoid wasting time/effort/money on something that won’t work. So if you can fail early and cheaply that is a great thing.

10. Make prototyping the cornerstone of your design process

You can prototype at each stage. Many people think of the flow as this:

Sketch -> Wireframe -> Mockup -> Prototype -> Development

Whereas it is better to think of it like this:

Sketch -> Prototype -> Wireframe -> Prototype -> Mockup -> Prototype -> Development

I hope that is as useful for you as the meeting was for me.

~ Dave

Jekyll is now running my live site.

2018-03-19T03:59:49+00:00

Goodbye WordPress - Hello Jekyll

I’ve been relatively happy with WordPress, especially since I haven’t posted very much in the past. I still am using it for some sites, but for this one, no more. Over the 3 day weekend for St. Patrick’s day, I finally finished moving the site between some other smaller work bits and some family time.

Why

So why did I bother moving? Well there are a few reasons listed here:

Terminal/Cli

I live in a terminal, and it’s pretty hard to update WordPress from vim. Now, I can make updating the blog part of my normal work flow, including editing it off-line. Also with git branching, I can have posts in progress. This let’s be just add small bits as I like the merge or PR to move a post live.

Patching/Maintenance

It’s kind of a hassle having to take the time and effort to keep a VM patched and running, along with updating WordPress itself as well as the plugins is kind of a pain when you got dozens of other hosts to worry about. Also there are security concerns if you don’t do this for WordPress. Now it’s GitHub’s problem 😜.

Simplicity

Still taling about the terminal here

I’ve been on a move to more and more simplicity where I can, I’ve started using Alpine for mail again. This is a terminal based mail client that is based on Pine which was my goto many years ago. I’ve also started using slack-term for my main work Slack team. I still use the full featured applications for some heavier tasks, but with tmux I can do all the things I need to with out taking my hands off the keyboard. Except to pick up my coffee cup.

Also with WordPress to me it felt like a production to make a post where as using vim is just putting some of my thoughts down over time then publishing with a git push. I think this will make it more likely for me to post updates. I have a number of posts I’ve been wanting to get out, but haven’t felt like going though the “ordeal” of it.

Cost

Well the small vps I was using isn’t much of a cost, but zero < any amount so there’s that.

What’s Next?

Well hopefully more posts (also hopefully useful/interesting). And of course I’ll be poking around Jekyll and seeing what tweaks I can make. Anyone seen any really nice sites built on Jekyll with something especially compelling?

~ Dave

Hello World! Moving to Jekyll

2018-02-10T23:59:49+00:00

Thanks to this post I was able to understand how to move from Wordpress (which had a mysql error by the way) to Jekyll on Github Pages in one evening while lying in bed … I should be sleeping now really.

Well I’ve added a little more here today (Sunday the 11th) but I haven’t decided if I’ll move the other posts, or more likely start fresh since most of my posts have either been reposted to LinkedIn and/or Medium. The real need I have is for some different landing pages I have for my social media profiles.

If you are curious you can see the code on GitHub. I can’t say it’s impressive, but if you are just looking in to jekyll it might be worth a look.

Update

As of 2018-02-11 19:20 dave-albert.com has come back from the dead! Now to run setup automated backups. I had them when the site was in Digital Ocean, but didn’t get to it when it was moved to Azure. Now will I keep it or move everything here? hmmmmmm…

2018-03-19 -> The site has been moved see here

Dublin DevOps meetup – Nginx (1 March 2017)

2017-03-01T23:59:49+00:00

On the evening of March 1st 2017 there was a great meetup event sponsored by Zendesk and Nginx (the company). The meet up was not surprisingly about the Nginx web server software. I wanted to relay some of the useful takeaways I got from the event. But before that I thought I’d give a very brief overview of Nginx and why I use it.

What is Nginx

For those of you unsure of what Nginx is (what are you reading this for then ?) or what makes it a compelling solution it is a highly performant web server. Nginx is similar to the Apache web server, but with a few main high level differences. I’m not going to dig in too deeply you can read more about that on this article from Digital Ocean or on this Quora Q&A.

For the record I use both Apache and Nginx, in fact if you are reading this the content has been served by both servers. Apache serves as a good web server for the WordPress blog, and Nginx is used for cacheing static content as well as a proxy/load-balancer. Here are the main points about each of the two different options from my point of view:

Apache Points

Good as a web server
Very simple to use as a php server
Slower serving static assets
Uses more resources (it has to load every module for every connection, so multiple site requirements can be very inefficient)
Can be used as a proxy, but depending on the setup as mentioned may be inefficient

Nginx Points

Can functions as a Load balancer, Proxy, Cache, and Webserver
Very fast, and can serve move simultaneous visitor on a single instance
Typically uses less resources to serve sites
For PHP additional services are required like (php-fpm)
If you are using a NodeJS server Nginx as a proxy in front of Node is a great solution

If you need more information on this feel free to ask me, and now on to the my key takeaways of the talks.

Improving API reliability with Nginx

by: Daniel Carmine
of: Zendesk

His talk was about how they improved the performance of their voice service accessing a third party providers service using Nginx.
oThe were operating a mixed voice and data service with a single set of Nginx proxies
They were dropping calls when Nginx was at a high load, regardless of the load on the voice service
They offered a service for their clients (many thousands of clients) that required uploading SSL/TLS certificates that would cause a reloading of Nginx services
Looking through their logs and analytics they realised reloading Nginx was causing the load on Nginx which in turn caused calls to drop
Once voice and data traffic were separated into different segments they found:
- Better voice metrics
- Better data metrics
- Better performance for voice
- Also better performance for data
Better to keep different types of traffic separate so it can be tuned for use This can seem obvious when laid out like this, but remember it’s quite compelling to use one solution to server both to save on resources and time.

Monitoring Nginx ingress controller on Kubernetes with Nginx Amplify

by: Michael Pleahakov
of: Nginx, Inc.

His talk was about using Kubernetes to manage a web stack that also managed the Nginx proxy container used for load balancing,

Kubernetes is for deploying and managing containers
Load balancing can be done with Kube Proxy for TCP connections
Kubernetes Ingres is for HTTP connections
Nginx allows for no downtime configuration refresh, which allows for better user experiance
He then performed a demo
- Showing how to use kubectl the Kubernetes command line tool
- Scaling is automated through ingress to Nginx
- It was a very nice demo, and if you are using Docker containers, I would recommend looking in to Kubernetes and Ingress
Ingress is an Nginx extension available on GitHub and also as a container on the DockerHub

After the Ingress portion of the talk we were given a brief overview of Nginx Amplify Nginx, Inc’s premium offering of a monitoring solution. Which I believe to be in a free beta period at this time (March 2017), so you can get a premium offering at no cost.

Amplify is a monitoring solution for the Nginx server by the Nginx company
A SaaS solution available from Nginx now in public beta
Demo
- Showing a nice looking dashboard with metrics of Nginx nodes
- It is available for Nginx and Nginx plus (the premium solution for Nginx)
- The analyser section of amplify shows an analysis of the current state and a static analysis of configuration settings
- There is also alerting available within Amplify
- I don’t know if I would consider sending my data out to a third party vendor, but it’s definitely something to keep an eye on.

You can find more from the company on Nginx.com/blog.

Advanced caching using Nginx and Lua

by: Cassiano Aquino
of: Zendesk

This presentation was about using Nnginx and the Lua programming language to imrpove a sites caching

Syshero.org
These slides should be in his GitHub /caquino.
Lua extensions can be compiled into Nginx
Lua (Portuguese for moon) programming language is embedded in many applications and tools
If you use Nginx with Lua you need to use the libraries designed to work with Lua to ensure the most efficient methods and they need to be synchronous
You can use Lua or luajit (Lua just in time compiler)
- Jit is more performant but has trade offs in stability and development speed
Basic caching is when you deliver assets (pages, images, css, js) from disk as quickly as possible as opposed to building and serving them through a CPU intensive server application
- Typically a cache is time based so that an asset is valid in the cache for 10 minute/hours/days/etc
- This is great when you either have a set schedule for deployments, minimal changes that get deployed, or have static content
- This is not so good when you have lots of changes, or constantly changing data
Header_filter_by_lua
- This is used to invalidate the cache on a POST so that the next GET will retrieve new content
The purpose of using Lua here is that the application doesn’t need to adjust headers or try to manage cache

Introducing Nginx plus

by: Owen Garrett
of: Nginx, Inc.

Nginx was initially develop by a single Russian developer trying to have a single server deliver 10,000 concurrent connections about 15 years ago
Non blocking event driven architecture
Nginx creates more consistency and efficiency
Nginx the company was started in 2011
Offices in San Francisco, Cork, Cambridge UK, and Moscow
Nginx surpassed usage of apache in January 2016
Nginx plus is the premium solution which adds value to the open source offering
The goal of plus is to respect the open source community by including additional features such as WAF, advanced load balancing, and enterprise features and development release as stable but released more frequently than the stable open source line as well as support
75% of the code pushed by Nginx, Inc engineers was pushed directly to the open source project not the plus offering

This was a good meetup, and I am definitely going to consider using the LUA option for cache invalidation for at least one of the projects we have going on at the moment. I’m not sure if Nginx Plus makes sense for me at the minute, but if I ever get to the point where full support for Nginx makes sense, I’d feel very good about having the engineers who are building it answer my calls?

~ Dave

Chatbots/AI meetup in Dublin

2017-02-16T23:59:49+00:00

CHAI Dublin meetup on 16-February-2017

This was about the second meetup of CHAI Dublin, and my first visit. It is a meetup about bringing product, UX, development and AI communities in Dublin closer together. I love bots, and want to see them grow in usage. They don’t solve all the problems out there, but they can solve many of them. Here are the points I took away from the meetup.

Building a Chatbot from Scratch

by Philip Kinlen

Philip was using IBM Watson Conversations tool to build his bot. The first 1000 API calls per month are free and $1 for each additional 400 API calls.

Intent -> Entity -> Dialog

which means:

What are you trying to respond for -> value that they supply -> information for the value

A simple example:

Hi -> where do you live? -> Dublin -> here’s the info for Dublin

For a simple yes no question you need to account for all of the following answer type:

Allow bail out
Unintelligible answers
Yes
No

A context would be something like someone’s name related to he or she in the chat.

It is very simple to get started and create a simple chatbot. They technology isn’t the hard part to getting started, it’s understanding your use case.

7 Things We Learned Developing HomeHero

Colm Moriarty

Building a tool to help people use energy more efficiently using image recognition.

“It is as easy and as hard as writing a book.” -> (It is relatively easy to build a bot, but is it any good?)
“Discovery is a challenge.” -> (How do people know they need it? How are they going to find it?)
“User expectations can be a problem.” -> (You are probably not as good as Alexa …yet, you need to manage their expectations.)
“You will be judged by the quality of your fallback.” -> (‘I don’t know what that means.’ That will irritate people, try to bring value even when the bot doesn’t understand what the user wants.)
“Even without bots, messaging is the future of customer service.” -> (Real time feedback matters!!!)
“Bots need hooks.” -> (Why are users going to continue using your bot?)
“We are still discovering what mobile devices are for.” -> (Are they even really phones any more? What are they really? What will they be?)

With bots, experiment and be bold!

Chatbot App Architect Overview

Foran Pavlovic

We already know what people expect from a web site, not exactly so from a bot.
Chatbots are a single entrance, but they can be messy since everyone interacts with them somewhat differently.
Bots don’t have the save level of understanding from a security perspective. We know how to prevent XSS/SQLi/etc, we don’t know all the new exposure point for bots.
Bots don’t always have the added context of imagery a website has, only text. That means language is an even more important consideration.
People will want bots, but we need to do a better job of UX when UI isn’t a factor.
Bots makes an asynchronous system much easier to use, making micro services even more compelling.
Messaging is still just an interface to our application.
Separating signal from noise is the hardest part.
We NEED a way to deal with angry/frustrated customers. Transfer to a live person as quickly as possible.
A bot should make your service better not worse.
Remember to include a menu/help/suggestions/etc.
Don’t make them think, instead let them choose.
Keep track of conversations, this will let you improve and help deal with concerns.
Be extremely careful with ambiguous questions, don’t try to play smart, instead be absolutely sure you’re doing what they want.

Conclusion

This was a GREAT meetup and will definitely be going again. If you are in Dublin and are interested in bots or AI I would highly recommend going to this groups events. If you want to talk about bots feel free to contact me.

~ Dave

And now for something new…

2016-12-16T23:59:49+00:00

Today was my last day in the office at Symantec, where for the last 2.5 years, I’ve been a Senior Principal Site Reliability Engineer for the Norton e-business team.

We’ve supported the current infrastructure, built tools and automations, performed a minor data centre migration, and started another major data centre migration.

This is basically what you might call “moving to the cloud”; or taking the servers that are currently in legacy data centres and moving them into a dynamic data centre (think AWS, Azure, etc.)

The reason I’ve left Symantec is to start up a new business with my co-founder, Julie O’ Donnell odonnellonline.ie. It’s called One15 and it’s a digital and marketing strategic consultancy firm. We also build digital solutions that we market ourselves.

Julie and I worked together at Publicis D Healthcare 3 years ago. She will be the CEO at One15 and I will be the CTO, or Chief Technical Officer, responsible for all things tech: the apps, websites, automations, chat-bots, servers, tools, tech team, technical strategy, etc.

Julie and I worked extremely well together at Publicis, where I was the lead developer and she was a company Board Director and Strategic Planning Director.

That’s how we know our relationship will be able to withstand the pressures and stresses of a start-up. We’ve both decided that working for other companies, and not building a long-term future, is not what we want to do.

We want to be able to work on the projects that we are passionate about, as opposed to working on something that is driven by someone other than ourselves. Obviously, we still work on projects for clients, and try to create solutions that are of value to our customers, but we want to have much more control over that.

When you work for someone else, there’s always uneasiness, at least for me; you never know what is driving the direction of the company and the action within your organisation. That is why our goal is to have more control over what we do day-to-day.

The great thing about the relationship Julie and I have forged is that I am very strong technically while her expertise lies in the marketing sphere. I understand infrastructure, software development, and how digital and marketing intersect from a technical perspective. Julie, on the other hand, understands how marketing and the digital world intersect from a marketing perspective. That makes it the perfect partnership.

Although I will be sad to no longer work with the people in my office and on my team globally from Symantec, I’m very excited about starting this new chapter of my life. I expect great things.

I know there will be struggles and that it will be harder some days than others, but I also know with a long-term view that we will do well because we are investing in our own future. On those days when it’s hard to get out of bed, it will be a bit easier remembering our ultimate goal.

Please get in touch with either Julie or myself at One15, if you’re in need of advice or assistance with anything digital or marketing related. We would love to help.

~ Dave

Agile & Entrepreneurship Meetup - Part 1: Jeff Gothelf

2016-02-02T23:59:49+00:00

"Shipped doesn't mean done" & other lessons any developer can apply today.

Recently I went to an “Agile & Entrepreneurship Dublin” Meetup featuring Jeff Gothelf, co-author of the best-selling Lean UX: Applying Lean Principles to Improve User Experience book, (via Skype) and Sean Finlay, who is Head of Propositions at Ammeon who spoke in person. Sean writes a DevOps and Lean-Agile blog. In this post I’m going to describe the group, the venue, the talks, and the key take aways I took to heart from the talks.

The Group (Agile & Entrepreneurship Dublin)

The meetup group was started by 3 colleagues of mine, and I wanted to support their efforts as well as the fact that I’m interested in the subject matter. The focus of the group is based on Agile/Lean/KanBan/Scrum software and business development, as well as topics of interest to the entrepreneurial community. The group was started on “Jul 20, 2015” and at this writing (31st January 2016) has 412 registered members.

“This Meetup group is for Agilistas, startups and innovators, who are interested in discussing how to fuse Agile with Entrepreneurship.

Come along to learn from top guest speakers, network, pitch for new job opportunities, have a beer and have some fun along the way.” - From their “about” on meetup.com

The Venue

Bank of Ireland on Grand Canal Square have a great looking space they offer to startups for meetings and as a work area. They kindly offer this space to the meetup group. As well as having some refreshments available including more wine and beer.

Good

Great looking space
Very nice people assisting with the setup
Top of the line A/V equipment
Convenient central location

Bad

Though my seat was comfortable since I arrived early, mostly they had stools that I heard some people were a bit dissatisfied with.
Unfortunately a pillar somewhat blocks the view of the screen for a portion of the room.

The Talks Introduction by A&E Organisers One of the most interesting things said in the welcome and introduction was that the organisers described themselves as experimentalists. Saying that the the venue, the Skype presentation, and even the meetup group itself were all experiments and they are testing, learning, and adjusting as they went along. I’d say we could all use a bit more of this attitude.

Jeff Gothelf

Ship, Sense, and Respond: A Culture Of Continuous Learning

Jeff gave a fantastic talk via Skype. The talk was so good in fact I hardly even noticed he wasn’t in the room. You can see much of the content of the talk in this video:

The biggest takeaways that really had meaning to me were as follows:

Create a culture of continuous learning (this is learning what your customer responds too, not just learning how to do something)
Shipping isn’t done. Done is when a meaningful change in customer behaviour happens.
Implement something and then study it. Was the change used? Was it used again? Are your customers more successful in using your product?
You can ship all the features in the world but they can still suck.
Continuous deployment is ultimately about continuous conversations with your customers.
Deliver is not the end goal, discovering emergent ideas and improvements.
Radical transparency, bring everything in to the light. Don’t hide things from each other, management or the customers.
The software is never finished so show it in its current state without fear. If we all know this is only a point in time for the constant evolvement of the product, then there is no need to be embarrassed by its current state.
Prioritise learning over delivery. It’s more important to learn more about what will help your customer become more successful using your product that to ship more features.
Maintain a position of humility, no one truly KNOWS the exact path forward so don’t act like you do.
Humility drives agility. This removes stagnation and allows you to be more responsive.
3 changes to make this a reality
- Change people’s motivation. Customer focused incentive structure (an iPhone app team is really the mobile commerce team) This requires objective measures of success (Did we change customer behaviour?)
- Change what gets funded (Evidence based funding decisions)
- Change how work is assigned. Give teams a problem to solve not a solution to implement (autonomy/transparency/learnings)
Process is a reflection of culture, if the process is broken you need to fix the culture first.
Shift culture of delivery to culture of learning

This may have been one of the most inspiring presentations I’ve been to in years (and I went to the Web Summit!). I’ll definitely be buying Jeff’s books. You can find out more of his upcoming book at sensingbook.com

You can read part to at: Agile & Entrepreneurship Meetup – Part 2: Seán Finlay

Let me know what you think, or if you have any questions.

~ Dave

Agile & Entrepreneurship Meetup - Part 2: Seán Finlay

2016-02-02T23:59:49+00:00

DevOps requires a culture shift and it's about people.

DevOps is indeed taking hold and Seán Finlay told us about how he sees it fitting into some organisations better than others.

This is part 2 you can find part 1 here: Agile & Entrepreneurship Meetup – Part 1: Jeff Gothelf

Seán Finlay

DevOps: taking hold

This was a solid introduction to what DevOps is, and what it means for an organization. It wasn’t deeply technical, but neither was the audience. Some of the takeaways from this talk that I thought had significant value were:

Driven by rapid delivery and delights customers faster, better, cheaper
The underpinnings to make it work are:
- Cultural shift
- Lean/agile learnings
- Automation
- Collaboration
- Cloud (Really this is more a tool that enables infrastructure deployment for smaller organisations)
One wouldn’t design a new business with a siloed organizational structure now
High performing organisations raise profits (puppet labs state of DevOps slide)
Focus on people first
Assess who you are and where you are
There are three type of organizations Pathological, Bureaucratic, and Generative (figure to the right). The Generative type of organisation it the one most capable of implementing DevOps properly
You have to fix the organisation for devops to work
Zero downtime upgrades are a major benefit allowing more frequent deployments including multiple times per day
Empower the team and create some early wins, this builds momentum and gets buy in from within the team and from around the organisation.

Ron Westrum's Type of Organisations

So I would say from both of these talks the most important thing if you want to move towards DevOps, then your culture and your people are what will make it work. There are lots of tools to enable DevOps processes, but they won’t make an organisation with the wrong culture work any better.

Let me know what you think, or if you have any questions.

~ Dave

Probable topics I will cover

2015-12-14T23:59:49+00:00

Some of the topics I plan to write about are related to software development and Linux systems administration. I have several technologies I find interesting, sometimes confusing, and often infuriating. I’ll be digging into plenty of different topics such as:

Docker
Ansible
RunDeck
CentOS
Ubuntu
Vagrant
NodeJS / JavaScript
Python
Functional programming
Elixir/Clojure/Haskell - These I want to make time for badly
Pebble watch, Raspberry Pi, Arduino, other microcontrollers
Agile/Lean
Audiobooks and podcast I find useful

If there’s anything specific you’d like to know more about let me know and I’ll try to find some value to add to the discussion.

~ Dave