Jackson Fish: Why I won’t hire people who want to be the “user advocate” on the development team.
What really struck me from the article was this:
I don’t care what your role is (i.e. developer, designer, tester, etc.). Everyone should be a user advocate. Everyone should have a common bar for the caliber and quality of the software the team is delivering.
It’s easy to have people advocating for users, it’s a lot harder to have a team advocating for the same user.
I remember catching myself wondering if this new ‘asset pipeline’ was worth having to learn about (great guide). Similarly, at first ‘bundler’ didn’t seem to work that well - now it seems like a no-brainer. Hoping to never get old :)
Facebook tech-talk notes
Yesterday I attended a tech talk at Facebook’s Seattle office. Kent Beck and two other really smart engineers (whose names I’ve forgotten!) talked about how they build software at Facebook. Really interesting stuff.
Some thoughts I sent to my team at Microsoft..
They too have a huge code base
Kent Beck opened by showing the diffs for the changes he made to introduce “civil union” and “domestic partnership” options for relationship status. There were like 15 files involved.. Yikes! This is not to say they have a huge, old codebase though. I was told at a previous talk that a given line of code stays around in the FB code-base for < 6months on average before it’s obsolete/refactored.
Huge emphasis on individual responsibility
Facebook engineers write their own code and their own tests. They own their code as long as it’s in the product and get paged at 2am when it breaks in production.
They have an incredible deployment / monitoring system
This is the main reason they’re able to deploy so often. They deploy internally constantly, then to 2% .. watching to make sure everything is ok.. then out to everyone. If something goes wrong — they notice that “Likes” are down 0.5% — they can roll it back for a region, for everyone, for certain types of users etc.
A culture of experimentation
There are some pretty famous examples of creative solutions like the HpHp compiler and pushing binaries out to production servers using torrents. What I found even more interesting though, is that both of these were only one of several completely different approaches to the problem that were built out either in full or to such an extent that the result became apparent.
Their deployment system also enables them to A/B test features constantly. Most UI changes get pushed out in several different versions and usage patterns dictate what ends up rolling to everyone. No relying on the stubborn opinion of Canadian PMs! ;-)
John Cleese on creativity
“It is easier to do trivial things that are urgent than important things that are not urgent” — John Cleese
Tech industry recruiting is laughably bad
I’m not recruited very often - indeed I get a little jealous of my friends who seem to be getting approached every other day by some exciting company or another. However, across my apparently desirable friends their recruiting experience is always the same - complete shit. I’d like to say that at Microsoft we do a fabulous job, but my experience on both sides of hiring decisions hasn’t been at all stellar.
In an environment where talent is in such demand, companies can’t afford to hire this core task out to clueless recruiting agencies.
Common complaints:
- Dropped correspondance
- Phone screen / interview out of step with expectations, wasting everyone’s time
- Candidates speak with one department on Monday and get cold called from another on Wednesday
- Conflicting answers to questions
Some fairly basic tips for recruiters:
- Use some fucking CRM software. Remember when someone says they’re only interested in jobs in NYC and don’t do front-end web development.
- Do what you say you say you’re going to - call on time and bring back the answers to questions you said you would.
- Have one point of contact.
- Be responsive.
- Don’t lie.
- Have a clue about the company and position you’re hiring for.
There is some hope though! Like travel agents before them, online tools are starting to replace incompetence and misaligned incentives with self-service. See whitetruffle, LinkedIn jobs, StackOverflow careers etc.
What I would do to save Sony
I just finished reading Hiroko Tabuchi’s excellent article How the Tech Parade Passed Sony By. I’ve felt bad for Sony for quite a while now. Despite having started out over 60 years ago with rice cookers and having successfully adapted to several huge technology shifts, Sony is on the verge of becoming irrelevant.
The article cites a few reasons for Sony’s decline:
- lack of focus - dozens of poorly differentiated models
- internal tension between the content and technology parts of the company
- failing to embrace trends in technology (digitization, a shift towards software, importance of the Internet)
I’d add to the list the following:
- many core products being commoditized (TVs, low end camcorders)
- insistence on proprietary formats and technology (ATRAC/mini disc, memory stick, etc) as the world has largely adopted open standards
If I were put in charge of Sony I would do everything I could to refocus the company on becoming a world class brand in home entertainment. A few ideas:
- Sell the Vaio business
- Get out of the phone / mobile audio / professional / camera / camcorder businesses
- Cut the TV unit to one model that comes in different sizes, all including a great video camera and supporting voice / gestural input
- Dramatically reduce the mess of other home theater products - there is no reason to be selling multiple mid-range receivers!
- Build a great software experience into the TV (apps for all major media / communication services, universal media queue, robot recommendations, share + recommend to your friends)
- Build great companion apps for every relevant platform
- Buy Netflix or Hulu if possible
- Offer all-you-can-eat access to Sony-owned content, for free (subject to release windows etc.
- Continue to fight hard for games.. I don’t really know about this market so can’t really comment
Thoughts?
PHP: a fractal of bad design
I haven’t used PHP in years admittedly, but in my first experiences with it even as an amateur developer it sure felt like something was off.
Between a thousand different, inconsistent ways of doing things and with the open source projects I encountered being hopelessly messy I quickly looked to something else.
Apologies to the Royal Canadian Mounted Police for the undoubtedly horrific pile of PHP I left them with.
Idea: Produce Spotting
More of a fun silly app than anything that could be reasonably be constructed into a business..
I’m in the middle of making a pizza and need some rosemary. I know there is a lot around our neighborhood but I can never remember where.
This app would let people record sources of wild, edible (not always?) plants and such.
Not sure that many would be eager to have their secret blackberry spot picked out.. maybe give something to get something?
Idea: flavors.me but for bars and restaurants
Restaurants and bars have really shitty web sites by far most of the time. The Oatmeal does a much better job of pointing out what they do wrong.
What does a restauranteur want out of a site?
- Key information (location, hours) to people already committed to going
- To be found and influence ‘where to go tonight?’ decisions
- Taking reservations (maybe)
- Engagement with their community (maybe)
- .. in a manner that makes their customers feel good about the experience
What does a potential consumer want from a web site? The very same stuff!
Dozens of services have stepped into this space (Yelp, OpenTable, Urbanspoon, Chowhound, Foodspotting etc etc) and are first stops for would-be restaurant goers. Consumers do not suffer the inability to find out about restaurants. However, through their abdication of the space, restaurants have lost control of the message. Potential customers see generic profile pages with *cell phone pictures of menus*.
Despite this, I still think there is value in a restaurant creating a reasonable web site. I’m sure tons of people search google for where to go to dinner, and as a consumer of nearly all of the above services I still find myself landing on these wretched restaurant pages frequently.
With so many restaurants being so poorly served by their web presence, I believe there is a business there.
What such a business might look like:
- Easy domain sign-up / transfer
- Great looking sites in seconds.. that actually work on mobile phones
- Painless menu management
- Upsell richer functionality like no-charge reservations, mailing lists, loyalty gifts, designable sites (ala Shopify), photography etc etc
Did I mention that in-restaurant software also sucks?
