There is tremendous learning value in interfacing a bot with a bot API directly, without going through an external bot building framework. To provide some context, the chatbot I built sends the user daily Twitter trends. You don’t need Twitter to use it. If you do have a Twitter account, it serves as a supplement to your Twitter experience, because the trends it sends you are country-specific. I learnt a lot about the interactions possible on a certain platform It is rather simple when the chatbot and the user can only communicate in plaintext. However, different chat platforms like Facebook Messenger, Kik or Telegram provide different features and interactions that change the user experience substantially. The best way to learn about what’s possible is to try their bot API directly, without going through an external framework or library. For example, Telegram has a unique inline query feature, where the user can simply type “@bot_name ..” from any chat to get some results. Therefore, the user does not always have to start a private chat with the bot in order to benefit from it. Once the query is submitted, the bot can process it and generate a list of results, which can be formatted in a variety of ways. When the user selects a result, the message will be sent to the chat. For performance reasons, lazy loading and caching are supported by Telegram. A sample list of inline query results As I was finishing my bot, I realized that it was not enough to send the user a list of Twitter trends. After all, the user may want to see tweets related to a certain trending topic, and one way to do that is to direct the user to Twitter via a specific link. Thankfully, I discovered that the Telegram API supports custom hyperlinks with HTML <a> tags, so I could keep the message really succinct! The user can simply tap on an individual trend to open up Twitter and view the tweets related to it. A list of trends with selectable hyperlinks In contrast, Facebook Messenger currently does not support custom hyperlinks. If I decide to implement the bot in Messenger, I may have to show both trends and their URLs separately, which will result in a very long message. As you can see, different platforms offer different features and interactions. The best way to learn about it is to try it. This may pay huge dividends in the long run when you need to optimize the user experience on different platforms. I learnt about service messages I thought that only the user could send messages to my bot. However, I was surprised to receive messages that were not directly sent by the user! These messages are known as service messages, which are administrative in nature. I realized that there was some value in handling some of these service messages. For example, when anyone (including my bot) joins a chat group, I will receive a service message from the platform indicating who it was. If it was my bot who had joined the group, I would send the introductory message as a reply without waiting for the user to say “/start”. Conversely, if my bot is evicted from a chat group, I will know about it and can treat that service message as a “/stop” message. I handled the different ways in which the user could stop my bot I informed the user to send “/stop” if they wanted to unsubscribe from daily updates. However, I learnt that not every user will send this message. The Telegram chat client allows the user to stop the bot by pressing and holding on the chat and then selecting “Delete and stop”. Unfortunately, I do not get a service message when this happens. When I attempt to send a daily update to this user, Telegram will reply with “403 Forbidden, Bot was blocked by the user”. I handled this error by removing the user from the list of subscribers, until the user decides to subscribe again. Therefore, I can save on future server resources as I will not try to send updates to this user anymore. Furthermore, it allows me to maintain an accurate count of the number of subscribers. What about frameworks? I have not tried using one yet. However, these are my initial thoughts on bot frameworks purely in terms of their interfacing capabilities with the APIs of various chat platforms. While there may be some initial time savings gained by using frameworks, they do introduce a layer of complexity and have various limitations. For example, since frameworks need to be rather general to cater to many use cases, they come with functions that you may not need. They may not support all API features of a particular platform as well. Without a framework, we can still build a bot that supports multiple platforms by having good architecture. Interfaces that parse received messages and send back a custom response need to be platform-specific, but other components such as the database and interpretation of text can be common modules. Thus, we have full control over how we want to use each platform’s API for an optimal user experience. , we can write just enough code for features that we want to use. software Furthermore Though there is more upfront cost, the code is still server-side. Hence, it is not as expensive to support two chat platforms as it is to support native Android and iOS apps. Just as with website builders and app builders that have come before bot builders, be aware of the pros and cons. Don’t feel obliged to use one! To conclude Part 1/2 In building my chatbot without the use of a bot framework, I learnt about the different interactions possible on the platform I chose, the different kinds of messages, and the different ways that a user could stop the bot. As a result, I have gained a much better sense of that platform’s true capabilities. If I do eventually use a cross-platform framework, I will have a better awareness of its limitations. In Part 2, I share more about parsing messages, handling timezones, and dealing with edge cases. Click here for ! Part 2 If you find this post insightful and want to see more, please ❤ and follow me!

Facebook

Twitter

What I learnt from building a chatbot without a framework (Part 1/2)

Logo by Pete Harris Hello. It’s been a while. Dear all, First: although I’m the technical cofounder of , I haven’t been a part of the community since 2009, and much smarter people than me have taken on core code management since then. This letter is meant with love, but I don’t presume to have any say. Elgg Still, I was forwarded , and while comments on that post are closed, I couldn’t not weigh in. In that post, Steve Clay says: Elgg at a crossroads Personally I really enjoy working with Elgg’s contributors and don’t want to stop, but for my career’s sake I can’t focus on reworking 2008-era APIs much longer. If those APIs are 2008-era, they’re also actually 2006-era. That’s a full decade of web technology ago. If I thought a global open source community was going to be holding to my API decisions a decade later, I might have thought harder about them. Elgg was born in another world. Our press called us “MySpace in a box”. ! Back then, a social platform had profiles, a feed, blogging, photos and tagging. To be sure, it was a simpler time. MySpace A decade later, we have Snapchat, Instagram, Periscope, and so on and so on. And Medium! More people access the web through a mobile device than anything else. Teenagers listen to music on YouTube. HD videoconferencing is the norm. We’re in the age of real-time context, mobile, bots, live video, and ubiquitous connectivity. There’s no good reason to keep backwards compatibility with the older era. That’s never been how the internet has worked. And the idea that you’d want plugins from 2008 to continue to work, or to continue to use Elgg 1.12, is ludicrous. I can’t speak for Dave Tosh, my cofounder, but I’m pretty sure he would have some choice words about that. I do too. I’ll leave them to your imagination. I started writing the first version of Elgg in 2003. Three years later, we decided we had to create a completely new version to progress, and we did — without backwards compatibility. There’s no reason not to do this again. Open source is amazing. If you want that stuff to work, you should fork the code and work on it yourself. Anyone can. (Confession: I thought about forking Elgg, too, six years or so ago.) do But the core developers should always be creating something new. They’ve given so much of their professional lives to the platform — it’s only fair that the platform be allowed to represent the best of what they can do. In turn, that’s a gift to the community, and to the web. (Another gift: the ability to make a living working on it. The comments to the post hinted at people who don’t think anyone should make money with open source, and I believe that attitude is incredibly harmful.) Please don’t let my ancient decisions hold you back. You should be having fun making new things that push the social web forward. And please accept my apologies for rudely wading in. Love, Ben Team Elgg, a decade ago, L-R: me, Dave Tosh, Misja Hoebe.

Harris

Instagram

YouTube

An open letter to Elgg

https://www.scalawilliam.com/1609/feature-switches-agile-scala-jmx/ Remember last time a critical service in production worked incorrectly and you reverted your changes, or restarted the services for new configuration? Neither do I. In one of my recent contracts, 10 days before production deployment I began to work on a feature that would actually take 2–3 months. The management had not accounted for the complexity of the task, but no worries: we have Agile and at our disposal. Play Scala Once you know that you’ll miss won’t make a deadline, all risk is gone, and only certainty is left. I agreed with the project manager to implement the simplest, dumbest, safest thing there was, so at least we would not miss the deadline. It would take 1 day to get into master. This would be our basic interface: And a very dumb implementation in , called ‘StaticFeature’ Scala Now knowing that the management will be asking for constant updates for the final deliverable, I agreed with the Product Owner that it’ll make more sense to implement something simple that at least partially meets the requirement, but would take a bit less. It took 2 weeks and we we were able to deploy it to production quite easily. If anything went wrong, we’d be able to switch back to something we agreed to earlier (ie ‘StaticFeature’). Welcome to ‘AdvancedFeature’, which is 10x more complicated than ‘StaticFeature’, is tested, and can fall back to ‘StaticFeature’ when it is unable to fit the bill. But let’s say rather than fail and fall back, ‘AdvancedFeature’ returned incorrect output? Because the input XML is very loose, there was no guarantee that everything would work as expected and none of it could be easily unit tested. However, how do we switch back? Revert Git code? Redeploy a configuration change? Come on! We have the JVM at our disposal. It has something called (“ ”) which provides you with a way to manage your application at runtime. You can do metrics, rebind ports, change configuration options, change logging verbosity, run diagnostics — all sorts of things. So here’s how you do it: JMX Java Management Extensions First you write a generic interface, a “Management Bean”: Then you write an implementation which registers this component to the JMX registry and allows its methods to be called remotely: And now when your app runs, you open up (jmc in UNIX shell), connect to the app, and then you can change values immediately (it’d call setFeatureLevel method upon pressing Return): Java Mission Control Of course in production you might want to have something easier to use, in which case I used to connect to a remote running process and change the value of configuration via a Shell script. Java 8’s JavaScript interpreter jjs Now the business had the certainty that instead of being stuck with incorrect behaviour they’ll get a lesser version which gives the expected behaviour. Final implementation, ‘ComplexFeature’: This was the most complicated by far, and the business wanted to do an A-B type of roll out of the feature into several production environments. Run on the first environment for a few weeks, then run it on the second, then the third, and then we’re done with the fourth. Should anything go wrong, we resort to the lesser versions and change the level at runtime. Easy & convenient. This approach was successful. Ops loved it. PO loved it. If feature C started producing incorrect results, it would fall back to B. If feature B started producing incorrect results, it would fall back to A. Ops were able to switch between these levels at runtime. Ops were able to deploy the same code to different production environments and A/B test the changes. Ops and Test were able to deploy the same code to production, staging and test environments and verify the behaviour of every single feature. Ops were able to monitor behaviour and the rates of fall back on DataDog. Something now is almost always better than nothing now. But everything now is better than something now, so choose Scala & JDK 8, it has it all. Connect up with me and . on LinkedIn Twitter — : A reader gave a very good question: Update, 4 Nov 2016 So what are the downsides and upsides compared to, say, database switches? My answer: No downsides compared to database switches, only upsides. The thing you want to work out is how to execute this trigger since your service is remote. 1. Java Mission Control via SSH tunnelling. You’ll need to specify extra Java properties to allow that.2. A custom script to make the call from the server itself. This is the approach I took in the article. Tooling takes a bit of time to build.3. Launch something like Hawtio on the server: This is a very nice simple approach but you’ll need to secure access control to it. Needs Jolokia: 4. Use with your own administrative panel via REST. Where you want to abstract the low level details. http://hawt.io/ https://jolokia.org/ https://jolokia.org/ Choose #1 if your developers have access to production machines and know what they’re doing (small team, few apps).Choose #2 if you DevOps and Dev are in close communication but Dev don’t have access to production.Choose #3 if you have more applications running on the machine, maybe even several instances of the same app.Choose #4 if you are delivering to a customer who has their own Ops team and need special reliable control of your stuff.I think this will warrant a further article.Databases shouldn’t really be used for configuration or management. But of course not all platforms support management facilities.

Oracle

Feature Switches, Inheritance and Agile with Scala & JMX on the JVM

Hey, my boss said to talk to you — I hear you know a lot about web authentication? -Yeah, I’m more of a passwordless authentication guy now. With cryptography being where it is right now, it’s hilarious how we used human memory for password authentication. I’m excited about how we can use SQRL and the blockchain for authentication now. It’s the future ! Cool. I’m just building a simple web app at the moment — a normal CRUD app using Rails, going to deploy to Heroku and use Auth0 for authentication with username and passwords to issue JWTs. Is that still the way to go? -Oh no. That’s old school. Passwords are dead. Securely authenticating with a side channel is the way to go! It’s the future! Oh, Okay. What’s that? -SQRL is a take on using a trusted device to authenticate you on websites. It bypasses the need to remember a weak human generated password. It’s the future! Squirrels? What? What do squirrels have to do with anything? -Not squirrels, SQRL! It sounds like squirrel but it’s spelled SQRL. SQRL. Okay, what is it? -SQRL shows you a prompt which you click to authenticate. It’s basically a nonce that your computer signs, or a QR code that you scan so that your phone signs it. The signed nonce is sent back to the server and you’re in! It’s the future! So like Oauth and logging in with Facebook? -No, Oauth is dead. Trusted third parties are dead! With SQRL you host the key on your own device! It’s the future! What? Wait, don’t users need to have a SQRL client installed on their devices to use it? -Well, yeah, but they can get SQRL on their phones or install a computer client. So people don’t have it on their devices? -No, but they’ll catch on. What? How? Who’s pushing the adoption for this? Google? -Oh, no, Gibson wrote the draft for this. Gibson? The guy in Brave Heart? -Not Mel Gibson, Steve Gibson! The guy in the Security Now! podcast. He made SpinRite, the disk recovery tool and said that Microsoft engineered a RCE backdoor in Windows. Microsoft engineered a backdoor in Windows?! What! -Oh, no, he was wrong about that. But then he wrote the draft for SQRL. Ummm, okay… so I need to add SQRL on my website? -Oh, no! SQRL is dead. Use Blockstack’s Auth JS to authenticate with the blockchain! It’s the future! Wait, what? What’s Blockstack. -It’s a distributed key-value store powered by a blockchain. Like the bitcoin Blockchain? -Yes, but no. Blockstack’s blockchain is for PKI, DNS and storing hashes. So Blockstack has it’s own blockchain? -No, it uses the bitcoin blockchain to store their virtual blockchain Like a side chain? -Yes, but it’s a virtual blockchain. Right, but what does this have to do with Authentication? -Blockstack’s Authentication looks up the user’s key from the blockchain. So the server doesn’t need to store that either! It’s trustless. It’s the future! But the user still needs to manage the private key! What if i need to log in from my dad’s computer? -Eventually someone will make a web based key store for private keys. Like Lastpass? -Yeah, but for private keys! It’ll be amazing! But then how will I log into that web based key store? -Maybe with a unique user identifier and a secret string. It’ll be amazing! I see. So I need to authenticate my web application with Blockstack Auth, which users don’t have a stable client for, yet. It’s like SQRL, but not SQRL because it uses a virtual blockchain to store the keys, which is actually a side chain. Users will be tied to their devices until LastPass starts storing private keys, and then I can log into my site from my dad’s computer by logging into LastPass ? with a username and password Yes! Isn’t it glorious?! I’m going with basic http-auth. Thank you for reading this! If you have time and like alpha things, check out this app I’m building: Mailpenny.com If you love distributed things, Blockstack is doing amazing work with their distributed DNS that may kill DNSSEC! Check them out at blockstack.org Big shoutout to Steve Gibson from for his podcast! It keeps me focused during my workout! Security Now! If you’re setting up authentication, I recommend to shave off weeks of development related to authentication. Their feature list is too long for me to list here. Auth0.com This piece was inspired by my musings with and this by CircleCI. Kobi post is how hackers start their afternoons. We’re a part of the family. We are now and happy to opportunities. Hacker Noon @AMI accepting submissions discuss advertising &sponsorship To learn more, , , or simply, read our about page like/message us on Facebook tweet/DM @HackerNoon. If you enjoyed this story, we recommend reading our and . Until next time, don’t take the realities of the world for granted! latest tech stories trending tech stories

Chain

Google

Heroku

Microsoft

Passwords are dead. It’s the future

There are a ton of for doing asynchronous effects in Redux apps. see this as a bad thing, but I love the intellectual curiosity and willingness to explore in the JavaScript community. libraries Some There wasn’t a effects that met my needs, so I ended up making something new. I’ll share the ideas I borrowed from other effects libraries and how they are used in . Redux library redux-funk For some background, there’s a that goes into why you’d want a tool for asynchronous effects in Redux apps. Stack Overflow post Redux Loop—Declarative effects in reducers The example below demonstrates some of the great ideas in Redux Loop. It’s code for a reducer that returns an instruction to increment a counter after a delay: Redux Loop Example The call to the function above can be read as “return the state unchanged and call with arguments [ ]. loop incrementAsync action.delay An advantage of putting async stuff into reducers is that there is now just one place to look when trying to figure out “ ”. Whether an action should be handled synchronously or asynchronously becomes an implementation detail, rather than a decision that effects what part of the app your code lives in. For comparison, with Redux Thunk, async stuff lives in action creators and with Redux Saga async stuff lives in sagas. What does the app do when this action is dispatched? A big rule in Redux is that reducers should be pure. It may initially seem that Redux Loop violates this rule, but all is well. Calling the function does not actually perform any side effect. Instead, returns a data structure representing the next state and effect. This is what makes it really easy to write tests—you can use deep equality checks to see what effects get returned: loop loop Test for a reducer that uses Redux Loop I didn’t use Redux Loop, because I needed to do custom reducer composition, such as delegating to another reducer from within a reducer. This is complicated with Redux Loop, because its effects using . must be passed up the state tree , , , and isLoop getEffect getModel batch The problem is that in reducer composition, we want the child reducers to only have access to part of the state tree, not the whole thing. So when storing effects on (like in Redux Loop), the effect somehow has to make its way to the top of the tree so that the middleware can see it. The alternative is to do an entire tree traversal, looking for effects, which probably isn’t a good idea. I borrowed a solution from Redux Side Effect and describe it in the next section. state Here’s the redux-funk version of the async counter reducer: Reducer using Redux Funk for declarative effects. I explain how is implemented in the next section. call In the code above, the is used to queue up an effect. The effect can be read as “call incrementAsync with action.delay as its first argument.” I use effects of the form . redux-funk adds a array at the top of the state tree to queue up async effects. So to test a reducer that uses redux-funk, you can just check the key: call [incrementAsync, [action.delay]] funks to refer to [func, [arguments…]] funks funks Redux Side Effect— Using the action object Redux Loop code has to sometimes jump through hoops to pass effects up the state tree. avoids those difficulties and has a because of a trick—it uses the object to communicate effects. Redux Side Effect tiny line count action The following code shows how you can add a side effect in a reducer using Redux Side Effect: Redux Side Effect example The Redux Side Effect middleware just checks the effects on the and calls them all with as the first argument. Using is a shortcut: while child reducers get only branches of the entire state tree, is shared across all reducers. action dispatch action action In spite of the elegance of this solution, I ended up not using Redux Side Effect because: The effects are non-declarative. In the example above, the only way to test that the delay is correct is to actually set a delay and compare timestamps. Reducers become dependent on the presence of the method. sideEffect Mutating the action object , which I’d like to avoid if possible. breaks the Redux paradigm redux-funk also uses the trick of storing actions on the object to avoid having to pass effects. But the implementation maintains the purity of reducers. adds the funk to an action, using a as a key: action call(action, funk) symbol Implementation of the function in redux-funk call When using redux-funk, you call on the top-level reducer, which takes the queue of effects from the action and puts them onto the store state. This restores the action to the way it was, and, more importantly, makes the funks available to store consumers. The testing example above used coalesceReducers coalesceEffects. Here’s the implementation of : coalesceEffects Implementation of coalesceFunks From React-Redux —Subscribing to the Store React-Redux contains helpers for using the state to render the UI, which is arguably the most important side effect. Most effects libraries for Redux are middleware, but, instead, React-Redux subscribes to the store and reacts to the current state. Since redux-funk makes effects as part of store state, you can just subscribe to the store and then program with the effects. I made a small helper, for handling effects in a way similar to how Redux Loop does it: runFunks, Helper for working with declarative effects enables the effects described in reducers to actually get called. In the counter example, the function returns a promise for an action. calls the function and holds onto the promise for the action. When the promise resolves, dispatches the action: runFunks runFunks INCREMENT runFunks Reducer using Redux Funk for declarative effects Here’s the implementation of runFunks: Implementation of runFunks You can use redux-funk without , and implement your own logic for handling. That’s one of the advantages of storing effects on For example, you could use a variant of the code above to add delays to every effect, debounce, use callbacks instead of promises, do remote procedure calls, inject dependencies into funks, etc. runFunks state. Try it Out If you’ve made it this far, you’ve seen the entire source code for redux-funk! Hope you get a chance and play with the , which are adapted from the . to try it out examples Redux Saga examples

HackerNoon Tech Stories Archive September 30th, 2016

What I learnt from building a chatbot without a framework (Part 1/2)

@kevinze
Kevin Lee

An open letter to Elgg

@benwerd
Ben Werdmuller

Feature Switches, Inheritance and Agile with Scala & JMX on the JVM

@ScalaWilliam
William Narmontas

Passwords are dead. It’s the future

@gilani
Amin Shah Gilani

Redux Funk—Simple, Testable Async Effects

@maxheiber
Max Heiber

HackerNoon Tech Stories Archive September 30th, 2016

What I learnt from building a chatbot without a framework (Part 1/2)

@kevinzeKevin Lee

An open letter to Elgg

@benwerdBen Werdmuller

Feature Switches, Inheritance and Agile with Scala & JMX on the JVM

@ScalaWilliamWilliam Narmontas

Passwords are dead. It’s the future

@gilaniAmin Shah Gilani

Redux Funk—Simple, Testable Async Effects

@maxheiberMax Heiber

@kevinze
Kevin Lee

@benwerd
Ben Werdmuller

@ScalaWilliam
William Narmontas

@gilani
Amin Shah Gilani

@maxheiber
Max Heiber