Some cracking stuff on there including Danny J Lewis, Sunship, and Jocelyn Brown…
I read a lot online about how chatbots are the future. Ultimately, the proof of chatbots is in the eating—I mean, err, talking (this really is a crap metaphor). Let’s try one out to do something like… booking a flight.
The “domain model” of commercial passenger aviation is familiar to most people, but still there are a lot of choices to be made. Which airline? Which day do I want to fly? Some days are a bit cheaper than others. What seat class do I want? What ticket class do I want? (They are different things.) Do I want to book using my frequent flyer miles or with cash? Do I get checked baggage or just hand baggage? It might be slightly cheaper if I fly from London Gatwick rather than Heathrow. Do I want to land at Newark or JFK? Am I logged into my frequent flyer account? Do I need to supply Advance Passenger Information now or later? What about if I need to change flights? If I book business class for the long-haul segment, am I going to get business class on the short-haul codeshare segment? What are the cancellation conditions?
Apps like those provided by British Airways do their best to hide away the complexity of this. Within the confines of “I know I want to fly with BA”, rather than “I want to compare different airlines”, the app from BA provides a pretty good experience at managing this matrix of different options.
Now let’s try and book a flight using a chatbot.
First stop, Facebook Messenger. After all, Mark Zuckerberg said:
We think that you should just be able to message a business in the same way you message a friend
Well, I know that I like messaging friends at 8:30 on a Saturday morning demanding that they look up flights for me. So let’s go!
I open up Facebook. ‘New Message’. I type in British Airways. And I get… nothing.
Okay, this is just the Messenger thing they build into the default Facebook view. I guess I should open up the actual Facebook Messenger page. Let’s have a chat with British Airways.
I found, well, British Airways i360. Which is a giant levitating dougnut on Brighton beach. I thought about messaging it and asking whether or not the doughnut is currently levitating or not. But then I realised that there’s probably someone being paid not very much money to answer stupid questions from people on the Internet and left it be.
Maybe British Airways haven’t climbed on board the chatbot train yet. I’m sure some enterprising developer must have built a chatbot that lets me search and book flights. Let’s search for one, on Facebook Messenger. I had an insightful conversation with ‘Flight Bot’ that resulted in me… not being able to search for a flight. That went well.
I mean, I get it. It seemed like a good idea when they built it, maybe not now.
I still want to find a chatbot that can help me fly from London Heathrow to Newark Liberty International. That shouldn’t be too hard. So, as with all problems, let’s turn to Google. Oh, Google, find me a travel chatbot.
Huzzah! There are plenty to choose from. The popular travel booking sites Expedia, Skyscanner, Kayak and Cheapflights all have chatbots. On Facebook! Okay, let’s test Expedia.
That didn’t work out. I’m now feeling guilty about wasting some poor human’s time with a test query when I could just search on the web. Let’s give Skyscanner a go. This time, instead of just being a chat window, it asks me to “Get Started”. This means it’s actually a bot, I guess.
The fact that it’s taken me this long to find a bot to help me book a flight, when the apps and websites are already to hand is… kind of telling. But whatever, let’s play along.
Hi Tom! 🙂 Welcome to the Skyscanner bot. We’ll help you find your next trip ✈. What would you like to do?
- Find a flight deal
- Tell me more
I guess I should “Find a flight deal”.
OK, let’s start again… where would you like to go?
“New York City”. If you can get me there, you can get me anywhere.
A flight from New York, United States
- Change origin city
- Set as home airport
Where are you going?
- Don’t know
- Show weekend trips
Err, no, hold on, you asked me where I would like to go. I’m in London right now. I want to go from London. Click, tap, London.
A flight from London, United Kingdom
- Change origin city
- Set as home airport
Where are you going?
“New York City.”
A flight from New York City.
- Change origin city
- Set as home airport
Where are you going?
- Don’t know
- Show weekend trips
Oh Jesus, I’ve managed to get this piece of shit to go into an infinite loop just by asking it to go from London to New York.
I’m not kidding.
I was going to write a piece about how chatbots necessarily have to provide an experience that doesn’t allow the user to control the details of a somewhat complicated, but ordinary, process like booking an international flight. That compared to using the website or the app, you’d be less able to control some of the complex matrix of choices you make when booking a flight. I was also going to point out that chatbots do not allow for comparison shopping in a way that, say, tabbed browsing or multiple windows does. Being able to open up three different airlines in three different tabs and search their websites to compare and contrast.
Instead, I ended up talking to a chatbot created by a major travel comparison site that failed to grasp the difference between the city I wished to depart from and my destination. I’d understand if this were a weekend hackathon project built by some people dosed up on energy drinks. But this is a product from one of the companies in the top five of comparison sites.
While faffing around with this ridiculous bot, I did get a nice message from a human at Expedia directing me to their website, their phone lines and wishing me a nice weekend. All of which is much more useful than a bloody chatbot that can’t decide whether I’m talking about whether I’m going to New York or from New York.
From this experience, chatbots are clearly the future of human-computer interaction. On behalf of programmers everywhere, I implore you to waste a few more billions on building them, even if they don’t actually help your customers achieve what they want to do.
🎵 Currently playing: LTJ Bukem - FABRICLIVE x Promo Mix 27.07.2010, with MC Conrad.
Still amazing after all these years.
The Chrome plugin Magnetify does what Spotify ought to already do: make it so it just opens the desktop app when you click a Spotify link in your browser.
The small army of browser extensions one has to install to keep using the web a tolerable experience is pretty telling.
URL Tracking Snipper is a Chrome extension that removes all those nasty
utm_source parameters from the query strings of URLs. Works very nicely. I already try to remove these parameters by hand, but automating the process is helpful.
Beyond being potentially creepiness-enabling, tracking parameters make things like bookmarks less useful. If I bookmark URL x, and then I bookmark URL x1 which is the same content as URL x, but has a load of tracking crap on the end of the URL, my bookmarking tools (Pinboard in my case) may not tell me I have already bookmarked it.
These parameters also misleading. If you post, say, a link with UTM parameters saying that the link came from Twitter and then I share that link via another channel, your analytics data now says you’ve got lots of people visiting via Twitter. But they didn’t come from Twitter—they came from someone who shared a link from Twitter in another channel. “Cool URIs don’t change” (said Tim Berners-Lee), but cool URIs also don’t replicate unnecessarily. In an ideal world, we’d have one URL for each piece of content.
These irritating query parameters make the lives of web users more frustrating solely to benefit somewhat creepy ad-tech people. Snapping these nasty protrusions off URLs benefits everyone. Except marketers. Oh well. I guess we’ll have to live with that.
🎵 Currently playing: The Funklovers Soulful In The Autumn Season 2011
🎵 Currently playing: Curtis Mayfield - Live At Ronnie Scott’s (1988)
On Monday, a man named Dev George posted the following on Mozilla’s Bugzilla bug reporting database:
Your notice of insecure password and/or log-in automatically appearing on the log-in for my website, Oil and Gas International is not wanted and was put there without our permission. Please remove it immediately. We have our own security system and it has never been breached in more than 15 years. Your notice is causing concern by our subscribers and is detrimental to our business.
What happens next is both completely predictable and an excellent illustration of how to not do security.
Browsers are in the process of making it so insecure logins are flagged up as a potential security issue. This recently went live in Firefox and is slated to go live in Chrome soon (it is available behind a feature flag).
The website referred to in the comment—Oil and Gas International (or O&GI)—not only accepts logins via HTTP, but also accepts credit card payments sent unencrypted. At least, it did. It is now offline.
Not only was the site not using Transport Layer Security for logins or card transactions, the site itself was rife with an array of fundamental security failures.
It was running on ASP.NET, version 2.0. .NET 2.0 was released in 2006. The version the OGI site was running was likely teeming with old, unpatched vulnerabilities. Microsoft stopped supporting .NET 2.0 a long time ago, and they no longer support versions of Windows capable of running .NET 2.0. I’m not an expert in .NET, but from what I understand, it is pretty trivial to upgrade the underlying .NET Framework (just as one can run old Java code on a newer JVM).
Not that framework vulnerabilities are the main problem. The site was vulnerable to a SQL injection attack in the login form. Before the site went down, people found that submitting poorly formatted username/passwords led to the .NET server responding with a stack trace with highlighted source code.2
Here’s another screenshot showing the application vomiting back stack traces when the login code fails to handle a
So far we’ve got no HTTPS, an extremely outdated back-end framework, SQL injection, and a server configured to reveal stack traces to the user.
According to this Reddit thread, someone took a peek at their user table and found all the passwords were stored in cleartext. And some charitable soul took it upon themselves to drop either the user table or the whole database (I can’t tell which). The Reddit thread also says that a user on there called the company and tried to explain their security failures, only to be rebuffed.
Nmap results posted in that thread and on Twitter showed that the server had open ports for Microsoft SQL Server, POP3, IMAP and… HTTPS. Yes, they had port 443 open for HTTPS traffic but didn’t use it.
At one level, the technical lessons of this are obvious. In no particular order:
I do have some sympathy for the site owner. According to this biographical page from (an archived version of) the site, he’s a writer with genuine expertise and understanding of his field, and has worked with some of the biggest companies in the business. He knows about oil rigs, not infosec.
He has every right to sell access to his expertise and knowledge through running a premium website. But doing so comes with risk. Information security risks are genuine risks that businesses have to cope with.
And they are doing a poor job of it. Every business that builds anything online needs to have a strategy for how to handle information security risks. None of the issues I raised above are complex. Every professional developer worth their salt should know the importance of getting them right. And most do.
As I’ve noted in the past, business organisations often talk up the complex, ever-changing, difficult nature of these threats. Listen to the non-technical discourse on information security and you’ll end up so confused as to basically do nothing. The threat is insidious, ever-mutating, confusing and impossible to fight that it is quite rational to not do anything. This ignores the truth: the threat model that for small and medium sized companies are less like Wikileaks and/or 24 and more like… well, this kind of small potatoes nonsense.
The threat to most small-to-medium sized companies is that they don’t seem to get this. Poorly built websites, often outsourced to the cheapest supplier, filled with utterly boring vulnerabilities. SQL injections and insecure form submissions aren’t the constant, dynamic, ever-changing threat that pundits in the press pontificate about. It’s not cyberwar, just run-of-the-mill bad coding, lack of testing, and poor systems administration.
In so many companies, people haven’t fully understood how bad software is. They haven’t grasped how widespread poor business information security truly is. They hide behind certifications and assurances.
I’ve seen organisations proudly show off their ISO 27001 certification, then taken a look at their site and found an old, unpatched POP3 server running without TLS and accepting passwords in cleartext. I’ve seen organisations offering to do ISO 27001 certification, as well as government-approved cybersecurity work, without HTTPS on their website.
I have seen so many people and organisations in business who think software is something that gets “done”. That you just finish writing the software and nothing else needs to be done. Everyone who works in software knows this isn’t true, but this incorrect attitude is worryingly frequent in business.
I have also seen organisations where non-technical staff are not instructed on how to deal with reports of security vulnerabilities. In such organisations, it’s only through luck that a vulnerability is passed on to developers/sysadmins who can actually do something about it. When a security vulnerability is revealed, the staff run around like headless chickens worrying about what to do. People at all levels, in all areas (sales, customer service, support, management etc.) will need to learn what to do when someone tells them their website or app has a vulnerability—what the appropriate reporting procedures are etc. They also need to have confidence that the business will take it seriously, and know how to get it fixed quickly and properly.
Those at the management level need to take responsibility for understanding some basic information security. Without basic knowledge of information security, they won’t be able to evaluate risks rationally. If people were wandering around so worried about laser-shooting stalker drones that they didn’t bother locking their front doors, that’d be a sign of a colossal misunderstanding of risk. People getting worried about three-letter agencies and cyberwar while not fixing (or even having basic awareness) SQL injections, XSS vulnerabilities and other well-understood threats is the situation we see now.
The same misplaced fears affect the wider public. People read stories in the press about, oh, the CIA, NSA, GCHQ, Snowden, Wikileaks, mass surveillance, and cyberwar. That’s all worth thinking about. But most people, most businesses, are unlikely to be the target of surveillance by a spooky government agency, nor are they likely to be pawns in a global cyberwar. Much more likely is some incompetent website designed by inexperienced, self-taught developers with some major Dunning–Kruger issues, unwilling to learn anything new, is going to accidentally leak private data due to, oh, the sort of important but boring sins committed by sites like O&GI.
The real problem with information security is management-related: those in charge, unable to grasp elementary security issues or evaluate their relevance, properly handle security reports (sometimes from scary pseudonymous people on the Internet), or listen to experts. The existing certification processes, and the attitude that security is about “reassuring” customers (rather than actually being secure) is part of the problem. O&GI is a case study in how bad technology choices and poor management combine to produce a predictable (if undeserved) outcome. Here’s the scary bit: the same people who design websites with this level of insecurity are also designing your internet-connected coffee machine or teddy bear or, oh, car. We can’t carry on without some pretty fundamental fixes to how information security is handled in society, in business and in government.
The following is contemporaneous notes from a talk at London Web Standards by Matt Bee (@mattbee) from futureLearn. There may be errors or material I missed.
Auto-didacticism: the act of self learning or self-education without the guidance of masters (such as teachers and professors) or institutions. (Audience member: “Just Stack Overflow!”)
Wholly self-taught front end developer. Two things happened when I first started developer: “The Facebook” and Firefox.
If you are self-taught, you are in good company: Leonard Da Vinci, Charles Darwin, Quentin Tarantino, David Bowie, Jimi Hendrix.
Our industry is amazing - the support/encouragement available is unbelievable. We are excluding new people though.
“I followed an unconventional route to being a web developer”. But that’s a lot of people.
Source for first website - table based layout, a lot of view source, a lot of Notepad, a lot of IE 6. Used to work mostly in HTML and CSS. With the help from books like “HTML for the World Wide Web - Visual Quickstart Guide”, learned a lot as a tinkerer.
Two years in: good with HTML (table layouts) and moderate CSS (fairly new), basic PHP, could use FTP and do basic web config. Could get a site up and running from scratch. This was enough to get my first developer job. This was without any computer science background.
I worry for the future. Not in a political sense either. The openness is being removed. A lot harder for someone who is intrigued to learn. 69.1% of developers are self-taught according to Stack Overflow.
3 key areas that have made it harder for new people:
My first computer: ZX Spectrum 48K. Had to load programs off a tape. Instruction manual contained programs you could write and run, could produce beautiful graphics (if you had a colour TV). You could buy magazines to copy commands.
Compare the ZX Spectrum keyboard (with “str$”, “rnd”, “peek”, “poke”) keys with the PS2 controller. UX won out - but blackbox devices don’t enable developers.
Terminal and command line are hidden away. Accessible hardware is around, but you need to know about it and have surplus cash for it. Simulated environments aren’t always great experiences.
It used to be easy. View source was enough to get inspiration and troubleshoot. A lot of code was not dependent on anything else. Generally you saw exactly what people write. Didn’t depend on a node package that the author of the code you are looking at didn’t understand. The was true even for big production sites.
Today: a new developer can create with no exposure to real code. View source is now potentially useless. Everything output is transformed from what is written. HTML and CSS sometimes feel like second class citizens.
Tweet from @paulcuth: “Just starting to use Web Components & Shadow DOM. Great fun, but going to raise the barrier to entry massively. View source is practically useless.”
Complex tools and mass adoption: “I first thought of learning AngularJS through their official documentation, but this turned out to be a terrible idea.” (source)
Try to install a Node.js library. Thousands of C++ compilation errors. Answer is to download a few gigabytes of Xcode that you probably won’t ever use. Why?2
We are exposing thousands of developers to this every day, and potentially putting them off a career in tech.
I am not saying do not use frameworks. They are valuable and save time and effort. I am saying provide a pathway to learn them.
In summary: I don’t think I’d cut it today. I’m not a great developer, I have huge gaps in my knowledge. I do have other attributes that make me valuable.
“The rapidness of getting the information doesn’t really correlate with the pace of the actual learning. This frustrates us” - Gavin Strange
What can we do? Consider the future developers. Run a code club. Run a Homebrew Website Club. Help teach people. Promote diversity on teams.
People learn at different rates and in different ways. Empathy is important. Provide pathways for great people to become great developers.
(All pictures from Wikimedia Commons!)
The following is contemporaneous notes from talk at London Web Standards by Matteo Pescarin (@ilPeach), development lead at Sainsbury’s (for now). There may be errors or material I missed.
“I’ve seen things you people wouldn’t believe”. I’ve seen developers and designers who didn’t want to touch a single line of HTML and CSS. I’ve seen fantastic designs that would work on paper. I’ve seen code that shouldn’t work in browsers.
There’s a lot to take in: learning the ever increasing complexity of front-end technologies.
At the beginning, we were very happy. Using tables to display data. Paginating millions of records. Then UX.
Then we started looking into content-first approaches. Personas, user journeys, interactions and behaviours, usability, cross-device compatibility, and many other aspects. A lot to take in.
Let’s say you don’t know web design. What is web design? Everything? Multi-faceted set of practices and doctrines that surround websites. Maybe it is more similar to product design. Quite difficult to grasp everything.
Semantic markup. Do you know?
If developers have too much to take in, what should designers do? Learn to code?
The developers are masters in their field. Developers know the limits of the technologies. Designers know colour theory, typography, grid systems, use of whitespace etc. “Don’t expect to find the unicorn” - there are developers who learn design, and designers who learn to become developers, but they are rare.
Don’t expect tools to solve your problems: tools exist to help developers do design better. Things like SketchUp (and earlier Dreamweaver) enable designers to turn ideas into web pages. Developers have the same: CSS grid systems taught us indirectly how design work. Methodologies like Atomic Design too. But don’t let anybody decide which tools you should be using. Allow yourself to be flexible enough to experiment, try new things.
A better world is possible. We cannot shift responsibility.
1-10-100 rule: if you have to fix a problem during design time, cost is 1. Fix a problem during development time, cost is 10. Fix after delivery, cost is 100. Data-based analysis showed the 100 is more like 1m.
Iterate and validate: as developers and designers, we are biased. Involve your users. Design in the browser. Rapid prototyping. Use Design Sprints. Use style guides and pattern libraries - save thousands of hours in the long run.
Designers must sit with developers, to promote knowledge sharing. Work bottom-up, understand each other, elevate developers to work on front-end design.
The following is contemporaneous notes from a talk given at London Web Standards by Imogen Levy (@teppie). There may be errors or material I missed.
Dementia: memory loss, mood changes, changes in reasoning. Caused by a number of diseases - most common is Alzheimer’s. Over 850,000 people in the United Kingdom with dementia. By 2021, over 1m people will live with dementia. Getting worse with ageing population.
Designing websites for people with dementia is an unexplored space. Redesigning the Alzheimer’s Society website, it was key that it worked for people with dementia, so they were included them at all stages in the design process.
First step was to develop a series of personas. Personas help guide decisions in design. What are their needs? Used focus groups, interviews, help line call logs, surveys and quantitative data from existing website. Focus groups included both people with dementia and carers. Focus groups run through local branches of Alzheimer’s Society as well as local charitable groups included singing groups, “Alzheimer’s Cafés” etc.
Challenge: “website is informative scary”. Contained 10,000 pages of content, but was added to without review. Overgrown, unclear structure. “Time to do some weeding”.
Google rankings were excellent, but navigating via the website wasn’t good. Old site started with four items on the left-hand navigation panel, but this grew. Had to start again, but didn’t want to lose out on Google rankings for key content.
New IA/site structure to make it easy for people to find stuff easily and quickly. Worked with consultants.
Digital team owned website, but dementia content was produced mostly by team responsible for print content including fact sheets and material handed out in doctors surgery. Printed material must meet NHS Information Standard.
Had to restructure content but not break compliance with NHS information standard.
The goal of the website is to help users find information and complete tasks.
Structure tested using online tool called TreeJack.
Set of tasks given to user, then user is asked to navigate through website. No design, just structure. Gives you a “spider graph”. The cleaner the graph, the easier the navigation.
Example task: find out more information about vascular dementia. 90% found information easily. Each click stage broken down.
Unsuccessful task: “how can you best look after yourself while caring for a relative?” Content was contained within the “About dementia” section, but most users went to the “Get Support” section, though that did not contain the relevant information. Result was moving the information.
Wanted a new brand strategy as part of a wider goal to change societal perceptions of dementia - “I have dementia, and I’m OK.”
Use of logo across multiple devices. Square logos on a mobile phone was hard.
Colour palette: web use was limited. Lots of testing required. Two prototypes tested - one wireframe - to test out IA hierarchy. Testers liked the design and layout, gave positive impression
ICO enforcement notice against AS. Data breaches including website hack. Great deal of scrutiny followed. Complete restructure necessitated a new environment. Involvement of lawyers, penetration testers and security companies. A great deal of security scrutiny was performed.
CMS limitations: 10,000 to 2,000 pages meant a lot of redirects. Significant number of spreadsheets sent for 301 redirects.
New site enabled “friendly URLs”. Not perfect, but better.
AS wants to reach every person who has a dementia diagnosis, and “build stuff people want”. Team wanted to show how to be user focussed.
Visual Studio Code really is great, but my one true love is and probably always will be Vim. I recently upgraded to Vim 8.0 which came out in September. The async I/O and channels support should make it so things like linters can be adapted to not block the editing thread on save. This is good.
Today, I finally cleared up my
.vimrc, refactored it out into some separate files, and took a position on the Vundle vs Pathogen debate. I’ve used Pathogen in the past, but Vundle seems preferable. Git submodules are an unending source of pain and confusion. Vundle lets me just specify a list of plugins and have that list stored in version control, much like a
Gemfile or a
bundle.json or whatever Node is doing this week to download
These are the packages I actually use lots in Vim:
.editorconfigand you put it in the root of your project. It specifies all the editor settings: character set, indent style, indent spacing, end-of-line coding, final newline etc.
node_modulesfolder that the front-end guys use” could be specified in a shared config file.
I have a bunch of other stuff installed in Vim including fugitive (Git integration), ack.vim (which gives you integration with the
ag search tools, which are basically faster versions of
grep) and vim-rails (no prizes for guessing what that does). I don’t tend to use them. Two reasons. One is that I will often just use the command line. I don’t really feel like I need to use Vim as a command line replacement. 2 For something like Git, I either use the command line, or I use SourceTree. I don’t really feel like I need it baked into my editor.
The second reason I don’t use plugins I have downloaded: I don’t use them just because I don’t use them. They haven’t stuck in my head. They are quite likely to be useful but I haven’t burned the commands into my mind like I have with Vim itself and with the other plugins I use. One of the things that Visual Studio Code, Sublime Text 23 and TextMate before it did was make commands discoverable quite a bit easier by having a commands “palette” you could use. I should probably have a personal Vim cheatsheet. I already know the basic functionality of Vim well enough but each plugin brings some potentially powerful new stuff that you don’t really bother using until you actually learn how to use it.
There’s a few things I want to try soon. One is setting up ale, the asynchronous linting engine for Vim, which will asynchronously run your code through a whole variety of linters for different languages. The most useful for me being
Oh, and while we are talking about linters, let’s talk about Proselint. Proselint is an actually useful linter for English prose. It isn’t a grammar checker. There are too many of those. I recently saw a grammar-checker-as-a-service called Grammarly being advertised at me during a YouTube video. The ad featured a rather clueless social media strategist needing a Chrome plugin to correctly spell basic words. The premium edition of Grammarly costs $139.95 a year. Jesus. Proselint is designed for people who are generally able to write in English but want the sort of thing one would consult a style guide over checked. I have a Proselint plugin setup in Visual Studio Code. Once I set up the ale plugin, I hope I will get the same in Vim.
I also tried vim-devicons and Powerline/Airline, but having to install custom fonts just wasn’t something I really felt like bothering with long term, so I stopped. I’m going to give Gundo a try soon.
Mako templates aren’t such a good idea, but Pyramid seems to use them. I tend to normally use either Jinja or Django templates, or ERb in Rails land (because HAML annoys me, and “logic-less templates” are a lie). ↩
I’m told. I have never really used any version of Sublime Text. After TextMate was left without updates for such a long time, I lost interest in ever using an editor that wasn’t both free-as-in-beer and free-as-in-Stallman. ↩
Jean Louis-Gassée has nailed the exact thing which the iPhone got right: wrestling control of the device away from the carriers.
This remains a problem in Android land. It isn’t a choice between empowering the device manufacturer or the user. The user doesn’t get the power–the network provider gets the power. And if the network provider doesn’t bother giving updates, the user gets all the vulnerabilities…
Bad news: Parliament voted down measures to make sex and relationship education required in British schools. Instead of comprehensive, rounded and LGBT-inclusive, British schools can continue to teach a minimal battery of reproductive biology combined with a tissue of religiously-inspired propaganda and call it education in order to satisfy the concerns of “faith communities”. In Britain not offending religious people is apparently a more important public policy outcome than teaching students that gay people are human, how to negotiate sexual consent and how to not get sexually transmitted diseases.
I’ve been reading @adactio’s new book. Pretty much all I have read is great, and I highly recommend reading it.
But in the grand spirit of pedantic blogging, I take issue with this:
That tend to is doing a lot of work.
I think Jeremy is conflating declarative with fault tolerant here. HTML and CSS are declarative and fault tolerant. SQL is declarative and decidedly fault intolerant (and quite hard to master to boot). PHP’s type system is a lot more permissive and tolerant (“weak”, one might say) than a language like Python. The former is great for newbies and non-programmers, because adding
1 + "1" (that is, adding a string and an integer together) will give you 2, or at least something that when printed to screen looks vaguely like a two, though under the covers it may be Cthulhu.1 And the behaviour of something like Python is great for jaded old gits like me who don’t want the system to magically convert strings into integers but to blow up as quickly as possible so that it can be fixed before all this nastiness gets stitched together with the rest of the code and causes some real major bugs. The same principle applies on a grander scale with stricter type systems like Java or the broad ML family (including things like Scala, F#, Haskell etc.).
Again, depends on the language. PHP is pretty permissive. If you make a very minor mistake, you can often just carry on and keep going. If you are feeling particularly careless about your errors, you prefix your expression with an @ sign and then security people get to laugh at you. I hesitate to say this was “designed” but it was at the very least chosen as a behaviour in PHP.
This error handling model, this permissiveness of web technologies, isn’t a side-effect of being declarative. It’s actually a property of them being designed for the web, of human choice by the creators. There is a cost to this. It has been incredibly hard for us to secure the web. Permissive error handling can and has enabled a whole class of security vulnerabilities.
If Postel’s Law gives you the ability to use
border-radius in your CSS or
aside in your HTML and some terrible old version of Internet Explorer happily ignoring it without vomiting XML warnings all across your screen, then Postel’s Law also comes with the cost of having to worry about downgrade attacks. We collectively left SSLv2 running long after it should have been dead and we got DROWN. We did the same with SSLv3 and we got POODLE. These are examples of ungraceful degradation and the sad cost is your server being vulnerable to being pwned.2
With the last few attacks on SSL/TLS, it wasn’t just nasty old versions of Internet Explorer on Windows XP getting shut out of the web, it was non-browser apps that talked to HTTPS-based APIs. The Logjam attack meant that a lot of people upgraded their servers to not serve DH keypairs that are below 1024-bit. For most current day browsers, this was not an issue. Apple, Mozilla, Google, Microsoft and others released patches for their current browsers. Java 6 didn’t get a patch for a very long time. If you had a Java desktop app that consumed an HTTPS-based RESTful API which had patched Logjam, that broke with no graceful degradation, and the general solution was to upgrade to a new version of Java. On OS X, Java used to be distributed by Apple, albeit as something of a reluctant afterthought. Upgrading every desktop Java user on OS X was a bit of a faff. (My particular Java Logjam issue was with… JOSM, the Java OpenStreetMap editor.)
Postel’s Law giveth and it taketh away. One could give this kind of rather pedantic counterexample to Jeremy’s account of Postel’s Law and then conclude hastily “given how badly the web is at solving these security problems, the grass on the native apps side of the fence might just be a little bit greener and yummier”. Oh dear. Just you wait. When every app on a platform is basically reimplementing a site-specific client, you might actually get more fragility. Consider our recent vulnerabilities with SSL/TLS. After something like Logjam, the bat signal went out: fix your servers, fix your clients. On the server side, generally it meant changing a config file for Apache or Nginx in a fairly trivial way and then restarting the web server process. On the client side, it meant downloading a patch for Chrome or Firefox or Safari or whatever. That may have just been rolled into the general system update (Safari) or rolled into the release cycle (Chrome) without the end user even being aware of it. The long tail3 of slightly oddball stuff like Java desktop apps, which tends to affect enterprise users, assorted weirdos like me, and niche use cases, took a bit longer to fix.
If every (client-server) app4 that could be in a browser, in the case of a security fail, fixing all those apps would be as simple as fixing the browser (and the server, but that’s a separate issue). If everything were a native app, you have to hope they are all using the system-level implementations of things like HTTP, TLS and JSON parsing, otherwise you have a hell of a job keeping them secure after vulnerabilities. We already see things going on in native-app-land (Napland?) that would cause a browser to throw a big stonking error: user data being sent in cleartext rather than over TLS being more common than I care to think about. But the native app won’t scream and shout and say “this is a bad, very wrong, no good idea, stop it unless you really really really want to”, because the UI was developed by the person who created the security issue to start with.
The downside to Postel’s Law5 is sometimes the graceful degradation is pretty damn graceless. Sadly, the least graceful degradation is often security-related. The web might still be better at mitigating those than all but the most attentive native app developers, or not. Browser manufacturers may be better at enforcing security policies retroactively than app stores, or they might not. We shall have to wait and see.
But we still love it.
Sausage typing: if it looks like a sausage, and tastes like a sausage, try not to think about what it really is. ↩
At risk of giving technical backing to rising reactionary movements, perhaps the modern day variation of Postel’s Law might be: “Be conservative in what you send, and be liberal in what you accept, to the degree that you can avoid catastrophic security failures.” ↩
Wow, it’s been a long time since we talked about them… ↩
Let’s leave aside the semantics of what is or isn’t an app. Incidentally, I was in a pub the other day and saw a few minutes of a football game. There was a big advert during the game from a major UK retail brand that said “DOWNLOAD THE [brand name] APP”. Why? I don’t know. We are back at the ghastly “VISIT OUR WEBSITE, WE WON’T TELL YOU WHY” stage with native mobile apps. I’m waiting for a soft drink brand to exhort me to download their app for no reason on the side of a bottle. ↩
I claim no originality in this observation. See here and here. Interestingly, if you take a look at RFC 760, where Postel’s Law is originally described, it has a rather less soundbitey remark just before it:
The implementation of a protocol must be robust. Each implementation must expect to interoperate with others created by different individuals. While the goal of this specification is to be explicit about the protocol there is the possibility of differing interpretations. In general, an implementation should be conservative in its sending behavior, and liberal in its receiving behavior. That is, it should be careful to send well-formed datagrams, but should accept any datagram that it can interpret (e.g., not object to technical errors where the meaning is still clear).
The first two sentences are key… ↩
There’s finally a reasonable looking podcasting client for Linux: Vocal. It has the things that iTunes (etc.) has had since the early days which everyone told me were unimportant like remembering where you left off.
Of course, now we have the Web Audio API, Electron and things like remoteStorage, we may even get cross-platform free software podcasting apps that sync what you have and haven’t listened to across devices. And then we’ll be back to where we were back in about 2008 with iTunes and the iPod…