I'm working on what I call "semantic version control". Namely, instead of diffing and merging text, the goal is to parse the program and diff/merge the AST.
EDIT: A better way to state the goal: I want to take advantage of the underlying structure of a computer program to inform the diff/merge. The first step just happens to be diffing ASTs, but I go beyond this.
I already have some neat features--for example, I can detect moves fairly robustly. That is, if you move a function, I can see that as a single action even if you modify the function a little bit when moving it (e.g. rename the parameters). Also, there is already more information for resolving conflicts in a three-way merge.
Unfortuntely, the performance is currently horrible, which makes it completely unusable. Once I fix that (I just need to memoize the diff function properly) and fiddle around with some of the algorithms a bit, I'll have something interesting (I hope :)). After that, I'll probably add a UI.
Also, I should add that I'm not the only person with this idea. A lot of the early design and features are based on YDiff[1], although I used no code from there. The merging and some of the other stuff I'm working on right now is more original though (I think).
Hey, glad you are working on this. I have later given a talk on structural version control together with demos of structural editing and ydiff. Here are the slides:
I think structural version control can open lots of possibilities that are not accessible by text-base version control systems. I'm not really focused on this area, but I'd be happen to to see more people getting interested in this and work on it.
As for the performance of diffing, I think the bottleneck is with detection of moving. That's something that text-based diff doesn't do at all, so not a big drawback. But with structural editors, those moves can be instantly detected and logically merged by the editors. That's too far into the future though. But I think structural diffing without move detection can be of good use already.
Another direction that's possible to detecting moving more efficiently is to use hashing, the same trick used to eliminate redundant expressions in a compiler.
I was just wondering what would be the use cases of this.
I mean to say code version control really is text version control. The whole idea is to track the changes, the cause for the change and who changed with some history. Added with this ability to do other code release related changes like revert, merge etc.
Right now it's just a better diff/merge tool. It's more accurate than a normal diff or merge, doesn't get confused by meaningless changes, can resolve more conflicts (theoretically) and should be able to produce a much better interactive merge experience.
I don't have any features beyond that right now--it does not actually manage different versions itself right now. Once I have the client finished, you will be able to use it along with a normal VCS like Git.
However, there are certain advantages to using an approach like this to manage commits as well as diffs and merges. For example, given the information I'm gathering, you would be able to automatically mark commits that only changed comments but not actual code. Additionally, you would be able to perform some fairly sophisticated queries on history--for example, you would be able to track a function across renames or just list all the names it ever had, which would be useful.
This is all possible, but writing a complete version control system is outside the scope of this side project. Additionally, it would be much easier to get users if this tool just augmented their existing workflow (e.g. Git) rather than trying to replace it altogether.
Hopefully this gives you a good idea of the sort of things I want to do. I have some other ideas as well, but some of them are probably too difficult to implement. I have to admit that even the relatively simple coffee I have so far turned out to be much more difficult to write than I exit expected.
Your question is actually very useful--it reminds me that I need to come up with a good elevator pitch for the system. I'll worry about it when I actually have a working prototype rather than a loose collection of slow functions :-).
The biggest benefit is more clear understanding of changes in development teams over significant periods of time. For example, when was this function changed? That's very hard to answer when the diffs are not semantic.
The Eclipse IDE provides semantic diffs on top of version control when viewing specific commits (pending or already committed) but this is obviously not part of the underlying version control system and is limited.
Well, I'm using the AST to power most of the analysis. So the first-order diff is lexical--it works on syntax ignoring whitespace and the like. The second-order stuff, like matching substructures, is based on the first-order diff, but is semantic.
I'm also planning to add more semantic features. In the short-term, I'd like to detect other simple refactorings (particularly renamed variables). In the long run, I might even add very language-specific features (like trying to use type information to help with merging).
I'm currently working on a little project that has a new spin on Pastebin. I'm playing with a Bayesian classifier (or attempting and learning a lot) to auto-detect the language and appropriately set the syntax highlighting automatically without a dropdown and a reload. The paste will also be saved automatically so there isn't a page load for anything.
I'm still hacking away and haven't got very far whilst I'm getting my head around the classifier. One other problem I've noticed is keeping track of ownership. One solution I have in mind is that there is no ownership: when it's modified it generates a new link to share with. It might be good for very simple code collaboration - I've tried doing this with Pastebin over Facebook chat and this should make the process a whole lot smoother.
I also need to think about how training will work. I envisioned it so that the user can train it i.e if it guesses wrong, the user would correct it by selecting the proper language and the program will learn for the future. I'd be putting a lot of faith in potential future users but I'm not expecting it to be big or anything, maybe just something me and my friends can share and work on code friction-free (some don't have GitHub/BitBucket/etc accounts, one doesn't like CAPTCHAs at all).
I'm currently going off 100% self-education here too so I'm having to get to grips with some other maths stuff like some symbols (such as | means conditional probability and _not_ to be confused logical OR), so if the project falls flat on its face at least I learned something from it :)
Instead of having users train the software, why not download and feed it a corpus of open source software? Writing a script that clones random Github/Bitbucket/Google Code projects and feeds the code to the trainer (using the file extension to get the language) should be easy enough.
That is a good idea. Might even be able to just get away with not displaying a language name or anything so if it's off a little bit, it won't matter much. For example, there'll be much less worry if some C# gets mistaken for Java in rare instances - I'm guessing in those instances the differences won't matter.
Why not just make it a branching model? Whenever a new user edits a paste it creates a branch off that paste. It allows seamless sharing, without links. Kinda like this: http://srctree.net/
Building HTML5 mobile app designer tools. Launched in late February, now ramen profitable, several YC rejections under our belt...we are going to bootstrap the damn thing!
I'm building a site that performs discounted cash flow analysis of stocks. Most of the existing work in this area comes from people selling custom spreadsheets. I think there's a good chance to disrupt that market, especially by combining it with education and a notification component.
You calculate the intrinsic value of the company by projecting how much real cash it can produce over its useful lifespan. Then you calculate the current value of that future money (you have a target rate of return you want to meet or exceed). That gives you a fair price per share right now.
It's a pretty standard Benjamin Graham valuation method.
In general, discounted cashflow means "if you will get $100 in 1 month i.e. future cashflow, what will be it be worth today i.e. discounted to today i.e. present value which should be little less than $100"
I recently deployed my own bootstrapped application called Mucho Support (http://muchosupport.com). It let's you build a contact list of team members and gives you a hotline (local phone number). When a business emergency occurs, a concerned party dials the number and Mucho Support will cycle through your list of supporters until you reach someone. You can also text your phone number and it will blast message everyone on your team with that message. It also emails everyone and tracks emergency activity.
I guess the idea is that automated systems are great in detecting hardware failure, but they lack the analytical skills your people have. "Crap we marked everything 80% off, instead of just one item" is a hard problem to automate around.
Only $10 a month for unlimited supporters (30 day trial). Keeping it simple for now as I drum up interest. Would love it if you guys would check it out.
I've also used this as a proxy for automated services like Pingdom. It texts your hotline, and Mucho Support blasts your entire team, not just one phone number.
I just finished building a widget, for beta testing, that startups can grab from my website that has a built in screen recording tool. So once they embed this on their own website they can entice users to record their interactions and it will even pick up on audio feedback as well.
You can see the widget on my site http://betapunch.com. The widget is the little punch icon in the right corner of the page. It's not great as far as design goes, but that's why I am working on the 2nd variation of the widget I will be launching in a couple days.
I am trying to get as many people to grab the code and put it on their startup websites so I can make sure display and functionality are all in order in various different browsers.
So if you have a moment and want to give it a shot for your site that would be great! You can remove it whenever you'd like, I just want to make sure it does what it's supposed to. :-)
Added to a wordpress theme footer in a few different places. Couldn't get it to work & doesn't look like there's any way to contact you via the betapunch site. Email in my profile if interested.
Good point. I know it's not crystal clear and that is largely my fault because I see it every day. I am hoping the newer style widget will clear things up maybe. The startup is allowed to customize an incentive when they get their widget code to display on their site. The purpose is to get something from the startup as a reward for beta testing their app. (A free upgraded plan, or some swag etc.)
Video recorded feedback has been proven to be much better than text based feedback. Why write out a problem and have something get lost in translation when you can show them and talk them through it?
Just finished a Perl app called Daystack. It's a web-based calendar and notebook that helps users manage their day and their ideas.
Right now I'm working on a seamless way to encrypt and decrypt notes in the browser. This way only the encrypted version of a user's note is on the server.
I'm working on a tool that helps you find the webpages you've previously come across.
This is primarily done by showing large screenshots with each bookmark.
I plan to add a variety of page-browsing tools to help you find that page quickly (from color palettes to cover-flow-esq flow).
If you could do that well, that would be amazing. There was a post years ago from a Mozilla intern on "Lifestream." It's a little bit like your idea, but I've wanted this for so long. Still haven't figured out how to do this, so I shall wait to see how yours turns out instead.
I personally think some of the ideas there are amazing. I attempted to contact the intern on whether there was any work still being done in this area, but received no response. I really hope someone can create a browser extension, or perhaps that Mozilla might still consider implementing this.
I currently have an obsessively organized library of bookmarks, which still fails me at times.
Edit: Is there any way to organize bookmarks within folders on your website? I see it is not possible in the bookmarklet.
I've been using Folindux for awhile now, and it's super useful for collecting inspiration websites. I did, at one time, have a Tumblr that did this, but this works so much better with the bookmarklet. It still has plenty of room to grow, and I'm excited to see where it goes from here.
Seems like a good idea, though I signed up and installed it and it's not working for me. I can't bookmark anything. I'm using Chrome for mac, version 19.0.1084.52
This simplification allows building a store by just using the html mockup from the designer and also allowed me to build a visual designer for the frontend.
It also has sankey diagram traffic visualization for better understanding user patterns, mcommerce and all the other modern requirements for ecommerce.
Still not ready for prime time but I hope to launch in a few months.
I'm working in a 2D RPG in ImpactJS. Spent the last week getting HTML 5 caching to work on desktop and iOS, then put the finishing touches on a bash script to update the cache manifest, compress all my JS files and rsync them to the server.
Love it. I'm not an engine guy, so having some of it done for me is wonderful. ImpactJS is well documented, reasonably fast, and has classes for the basics that work well. Plus it has a level editor that now includes collision detection. I'm satisfied with the $100 price.
Pay for performance SEO quoting system. You don't usually know what you're getting w/ SEO because it's billed per hour or in the typical standard, pro, enterprise packages that work well for web apps but not keywords w/ different competition levels.
And, it feels like the interests of firms aren't always tied to their clients. "Of course we have a guarantee...you're on the first page (of Ask.com)!"
So, I'm working on a quoting / billing structure where customer payments are dictated by the ranking improvements. Feedback so far is good, but of course, it's a challenge.
Since good SEO is a personal relationship with the client, it might benefit you to personalize your About Page. Clearly the price is right, but to sell me, I want to know that there is a team behind the effort.
I would also suggest a sales video explaining how much better performance based SEO is than the standard company. Put it where your screenshot is on the homepeage. Videos really connect with people, especially if you can get across your passion to really help your clients. something simple would cost you under a thousand if you found the right person to do it. (check out www.quiry.com "manifesto" or even go with some basic kinetic typography with music and your voice in background)
I'm building a web application for stormwater engineering that integrates local jurisdictional requirements of computation, design and reporting with industry-specific project management.
An open source WYSIWYG CMS and plugin platform running on Node.js and CoffeeScript, along with a hosting service and simple virtual hosting control panel.
The main advantages over other similar systems are:
* open source: MIT license
* ease of use: everything is a widget, including layout/design aspects
* efficiency: virtual hosting for many sites on one server/VPS, Node.js, caching, reduced requests etc.
* easy to extend: just enter Github user and repo name and hit publish to add a plugin
Right now I have tools/widgets for adding text (with lots of Google fonts), adding images, email address collection, pages (tabs), editing code, installing plugins, managing files, and a button to play Groove Salad. I am planning on finishing the EtherCalc (collaborative spreadsheet) widget, the EtherPad widget, and then recording a video.
The video will demo all of those things above and also show me quickly coding (with the code editor) and publishing a widget for a live updating collaborative data list that takes advantage of the built-in NowJS and Mongolian Deadbeef libraries.
When that stuff is ready I will start trying to raise money on KickStarter, and hopefully will have some funds to tighten up some basic things like security and hosting features. Then when its sort of Beta ready I will try to publicize the github repo again. I gave it out before but its really not ready to share the code at this point, very messy and pre-alpha.
Anyway I am interested to know what people think of the concept or my plan.
I feel like I always need more than one project on my plate, so I'm working on two main apps right now:
CloudPlay (http://cloudplay.fm), my first Mac app, is a music player that searches for music locally (iTunes) and online (YouTube, SoundCloud), similar to Spotlight, and makes it easy to build playlists and share them. I'm going to iterate on it lots this summer while in the nReduce (http://nreduce.com/) program.
Dialoggs (http://dialog.gs) is an app I'm building with my buddy Drew Wilson (of Pictos fame). It's like a topical Twitter with privacy. Dialoggs lets you create (or be suggested to you after a number of replies) discussion topics (known as Dialoggs) that can be public, invite-only, or completely private, making them great for collaboration within a team. We're working on UX refinements as we learn how people use the app, and sometime this summer, will have a more public launch. More info on my blog: http://amirmalik.net/2012/04/27/introducing-dialoggs
I'm working on a perl script that will serve as a migration assistant from WhatsUp Gold v8 to Nagios v3. My company has 1,300+ hosts and 2,500+ services, all in various groups and states of health. It is my job to port over everything.
The version of WhatsUp we are using does not have a good way to export the configuration, and no API to query, so I am basically screen-scraping the web GUI for all the relevant info. I then do some basic sanitization (there are some illegal chars in Nagios configs), error-checking (how to handle a URL with a loopback IP address), and some cross-reference (putting hosts in groups and attaching services to them).
Overall it has been a great learning experience for me. I got to know Nagios quite well, practiced my perl a bit, and got to use git beyond just the tutorials I had read.
I plan on open-sourcing this as soon as I get permission from my employer. Company time and all that.
I've really been liking the Mac App Store. Great potential for selling apps as an independent developer, and not nearly as saturated at the iPhone/Android app stores. Such as...
It was a really frustrating process getting it accepted actually. Took over 3 weeks and several rejections. Really depends on which reviewer you get I think.
Not very exciting, but I'm experimenting with a site to list sales tax rates by city: http://www.sale-tax.com. It's been fun to play with scraping government forms, SEO, and adsense. Maybe next year around tax time it will even be useful to people :)
It would be more useful if it was easy to compare places. For example, do I save money crossing the state line to buy something? Some states also charge sales tax on specific things that other states disregard (e.g., clothing).
I'm working on step 2 of a naming company (we create names for new startups and products). The plan is to offer a package with a name, logo, and business card, so I'm working on my logo design skills and bringing in some more designers.
Also, I'm building an copy editing service that's more transparent and straightforward. Most editors have complicated pricing schemes that vary depending on the topic, document type and whether they charge by the hour or by the number of words. Instead, I'm charging a flat fee per word, no matter what the project, and you'll be able to customize other features (like delivery date, format, etc.)
Hoping to get both startups rolling by the end of June.
"Nom" is the Latin root for "name" in pretty much every European language (Nom in French, Nome in Italian, Nombre in Spanish, etc.) and a villa is an upper class country house, so together, it's "Naming House".
I'm building an online workspace for uni students to collaborate on assignments and includes a calendar, contacts management with chat across platforms and an email client with document editor built in for pdf & doc.
I am working on a content publishing app for email in the ESP space called Sendicate http://sendicate.net. If you have ever battled with the ubiquitous Mailchimp/Constant Contact/etc etc then you definitely understand the pain. I have just started taking closed-beta sign ups and am aiming to get the first test customers on the app next week.
For those interested in a little more details I have started blogging about some of the ideas and concepts behind Sendicate at my blog at http://www.chadyj.com
I am working on a webapp called Vanish Drive. It lets you share files in the Cloud in a non persistent way. Get the sharing done and the files will be erased from the Cloud after some while.
Last week I re-structured the app (updated with my knowledge, because it's in the works for quite a while now) and yesterday I modified Backbone.js' Sync function to handle the CRUD actions via websockets.
I'm building an interpreter inspired by interactive fictions (IF) by infocom. The goal is to be able to generate your css/less (including a generated styleguide) and HTML page skeleton just by writing plain english.
Here is an example :
"btn can be a button element"
"btn can be a anchor element"
"large-button is a kind of btn"
"large-button color is black"
"large-button background-color is lighten(black, 10%)"
I'm still figuring out the language but i'm feeling that it can be really promising.
It's a sports team management site. Allows team captains/coaches to easily manage their schedules, stats, and players. I have a good number of teams using it, so it you know anybody who has a team send them a link.
I also just released my first iPhone app called Bullpen:
Trying to help find, attend and share what is exciting lately. The sources are completely open :-) I have facilitated private events and closed groups to share comments and photos. I did this because I believe proper socializing happens in real world and the place you socialize matters a lot.
So far we've had a bumpy ride but I am commited to make it work to build a rather truly social planet!
I'm slowly putting together a PC game called Moonman, inspired by Knytt Stories, Terraria, The Incredible Machine, and Dwarf Fortress .. when I'm not distracted reading HN of course. ;)
I'm trying to learn stenography, using Plover (open-source steno software: http://plover.stenoknight.com/ ) and a $40 gaming keyboard.
A steno keyboard has 20 buttons and you can press any number of them at a time (chorded). You write syllables and it unambiguously resolves your strokes into English text. Experienced stenographers can get 220 WPM that way.
I'm finishing up an ebook. I've found that startup founders struggle to tackle privacy issues. I'm hoping this ebook helps early stage startups find both a reason and a method to tackle privacy issues. If anyone is interested send me an email (in profile) and I'll send along a full copy -- I'd love to get your feedback.
I'm writing a photo editor using Qt (for future cross platform functionality) and OpenCL for GPU accelerated filters. It isn't exactly going to change the world but it's a fun side project to keep my brain from rotting in the cube farm from 9 - 5!
It's currently available for Windows (Linux is coming next) and is completely free.
I'm building a mobile app for my podcast listening website, http://flapcast.com (which is getting some pretty decent traction but probably isn't taking off since I don't have a mobile app). I can't wait to be able to listen to podcasts on my phone, hit pause, then continue listening to them on my computer :)
I'm working on a project to help researchers and medical doctors to easily stay up to date with state of the art research publications (1M+ publications/year, more than 2k journals), by categorizing and ranking papers.
We (me and the business co-founder, a student in medicine) should "go live" in about two weeks.
I'm building a virtual assistant that prioritizes your reading list. I've been building it for awhile and we made it to MassChallenge semi-finals, but no farther. Hopefully we'll launch soon :). If anyone here wants a beta invite, you can sign up at http://foldr.co (just click launch)
I'm making an app to create json apis from any publicly available small datasets and share them with others. Think of imdb top 250 movies, wikipedia list of tv show episodes, periodic table etc. I'm still working on making it fast. For preview check out http://apify.heroku.com
I'm working on a simple app "Smileship". I am new and trying to learn php/mysql while building this. My app is fairly simple. The main aim of my app is to inspire people to smile.
I'm building a book recommendation engine. You rate books you have read and we recommend other books you might like. We don't care about anything that doesn't improve the quality of the recommendations, so no book clubs, no author blogs, no user book lists...
http://wikisherpa.com/ - online/offline Wikitravel/OpenStreetMap mash-up travel-guide app. Been picking at it part-time for a couple of years now, have about 12,000 active users split roughly evenly between Android/iOS, am trying to work out what to do with it next.
A project manager for the low-budget film industry. It imports a script and generates the production board, call sheets, and budget. Hopefully saves filmmakers time and money and lets them focus more on the creative aspects of production.
I've been futzing around a lot with Hadoop since you last mentioned this on here and haven't been doing much SQL work. Any news on the development front? Anything new I should take a look at?
No real enhancements to report but some small bug fixes here and there. I am working on adding some reporting feature that will generate PDFs. I will certainly let you know when that is complete (even though I am thinking you are not exactly in that target group).
I am working on a private Android app provisioning service with the ability to silently update apps on rooted devices. It started as a way to easily deploy Android apps I develop to my multiple test devices and has become a great method to manage and update my volunteer beta testers.
I'm working on a Video Player as a service. Basically you host the video files yourself in your server/s3/cdn whatever. And you get a player with many features like captions,stats (maybe realtime), VAST2.0 ads, playlists, multi playlists and many more. Anyone like what you read?
A brain____ macro language and interpreter[0]. I've always really liked Brain____ and I wanted to make it easier to program in, and this is my attempt at that (still pretty early release, kinda buggy).
Playing (mainly - tuning at this moment) with a new book recommendation engine, http://dozen.softover.com
The main challenge/goal here - better than existing big players quality with much less data and resources
Over the weekend I built a little app to simplify selling on craigslist. It essentially turns your listing into an auction, tracks views and handles all of the inbound e-mails. All you have to do is accept the best offer(s) if you decide to accept any at all.
http://muxamp.com Muxamp is a YouTube and Soundcloud unified playlist I'm wrote to learn JavaScript, jQuery, and Node.JS. It interfaces with the YT and SC APIs to deliver a responsive, interactive playlist.
An internal (for now) local web app that provides 1-click installs for our numerous Github-hosted projects, applications, and Wordpress blogs using Vagrant, Capistrano and PhantomJS (environment, deployment, and change notifications, respectively).
Working on the next version of http://ligues.ca (sports leagues manager). Also working on http://playffs.com (NHL playoffs pool) until the end of NHL playoffs.
Building a Pinterest for Instagram called, wait for it, Pinstagram http://www.pinstagram.co We launched it on HN less than a month ago and are getting some users so we'll keep working on it.
I'm working on my personal tracking app, RoutineTap. http://routinetap.com I'm having a hard time getting visitors; right now I'm also learning backbone.js and writing a mobile app.
I'm working on the marketing/traction strategy for my second startup.
Lack of marketing was the reason of the fall of my first startup... so I don't want to get burn a second time. I'll share some stuff on HN in a few weeks :)
Non-programming wise, I'm finishing up my PhD thesis. And my current side project, which I'll start as soon as I'm done writing is a small location based iPhone app.
I'm building a web package in Lua/Kepler. The idea is that you install it, create an admin account, and then decide what you want the site to be. If you want a wiki, then it installs wikimedia and configures all that. If you want a blog, it'll install a custom blog software with lots of extensibility. If you want CMS, then it will install Joomla/Drupal/Whatever, and you can configure all of that. If you want forums, then you can install phpbb/vbullshit/custom forum software.
Right now, I'm mostly designing blogging software and forum software. I plan on making everything first, releasing it all separately, then just packing it all up as a chameleon package once everything is finished.
I'm writing a little Sinatra app for my fiancee's birthday. It's nothing complicated, it just displays a different reason I love her each time the page loads (or you can permalink to a specific reason).
OK , as music lover (mostly as listener ) i developed weird abstract product that collect music recommendations from your Facebook friends and groups into private database , and in real time and let you play it ( you tube streaming )
its GUI sucks and its desktop app. but from programmer point of view it is doing its job and there are few weird people that like it :
http://www.tunesamp.com/
hu .. and i very open to suggestion how to develop it or which direction to take it
EDIT: A better way to state the goal: I want to take advantage of the underlying structure of a computer program to inform the diff/merge. The first step just happens to be diffing ASTs, but I go beyond this.
I already have some neat features--for example, I can detect moves fairly robustly. That is, if you move a function, I can see that as a single action even if you modify the function a little bit when moving it (e.g. rename the parameters). Also, there is already more information for resolving conflicts in a three-way merge.
Unfortuntely, the performance is currently horrible, which makes it completely unusable. Once I fix that (I just need to memoize the diff function properly) and fiddle around with some of the algorithms a bit, I'll have something interesting (I hope :)). After that, I'll probably add a UI.
Also, I should add that I'm not the only person with this idea. A lot of the early design and features are based on YDiff[1], although I used no code from there. The merging and some of the other stuff I'm working on right now is more original though (I think).
[1]: http://yinwang0.wordpress.com/2012/01/03/ydiff/