Client's print server crashed in the middle of the night. Month end accounting batch (written by idiots) updated databases before encountering dead end at printing. Accounting wants their reports, but there aren't any. Can't restore because sales & fulfillment have already done many database updates before I was called. Not sure what to do.
Today I was going to add a few features to my web app. Instead, I guess I'll be changing 20 year old programs to not update databases, but print reports. I think. Not sure.
So I thought I would just have a cup of Columbian Supreme, a donut, and a quick visit here. Makes about as much sense as anything else.
I'm working on an online comedy troupe. Basically, there will be a blog that a cast of 7 contributors have authorship access to. All of their contributions are rated by the audience, and a the end of the week, the lowest scoring member gets axed, while a high performing member of the community gets promoted.
- A few months into a new start-up in the mobile space, Kimua, working towards having a prototype by the summer (my partner and I have hired two programmers, so I don't have to do any of the coding here!):
- Doing some heavy research in functional programming for my work-in-progress text editor (codename 'ngedit', the final actual name will probably be 'kodumi'):
I'm going to restrict the number of hours per week I spend on freelancing and raise my prices as I reach that cap of hours. Hopefully supply is plentiful enough for this to work. Has anyone tried something like this? How did it work out?
I'm building better MRI reconstruction software. Most existing software does a filtered FFT + standard image processing. This throws away information solely because radiologists and other humans can't use it directly.
I'm trying to extract that information and turn it into something usable by humans.
I know it's not exactly a Hacker Startup, but my wife and I have been building www.realphotography.com steadily over the past year
We've tried to leverage the internet and digital media as much as possible: the business has a blog, the vast majority of our promotion is on the internet, we sell the digital files, all proofing is done online. But I'd happily solicit any feedback from the resident hackers as to how we could do this even better.
And I'd like to head off the comments about Flash, yes we're aware of the disadvantages, and you'll see we do off a non-flash site for weddings. ;)
I'm working on a photo hosting service, http://Simplebucket.com
Basically, I relaunched the service on March 20th with a new interface, look and a bunch of other features. Now I'm working on a feature "Beeps" for our next release. "Beeps" is the social part of Simplebucket. That's all I can say for now.
PROPOSED SOLUTION: (1) Take a raw search feed; (2) Give searchers a tool that enables them to maintain a persistent search state (i.e., it keeps track of your previous queries and all the results you have clicked on while using the tool); (3) Use results from user data to aid in constructing a better ranking of the raw search feed.
At my day job: Just finished a django checkout/cart system for one of our websites (www.bigfootsnowshoes.com) and working on a new site and more Python/Django apps. Been using Python full time for the past 3 months and loving it. I hate fixing/updating old PHP code now :(
At night: A service that will help locals find what they want more easily for things like going out for a fun night, events, etc. Think a better Yellowpages. Where I am, nothing like this exists so it's something I will be using for myself as well. Plus it's a fun learning
And just as I was writing this post I got my Python in a nutshell book on my desk. Yay :)
I'm working on a small project related to baby naming which is a toy to get me familiar with RoR and experiment with building a registrationless user experience.
It's just an evening project and should be online later this week on Heroku.
A more trustworthy, controlled reviews and question-answer site geared towards preexisting mailing lists like those in big companies: http://www.askurpals.com/
You are right about the name but big company mailing lists is just starting point for us - mainly because they are easy to target. In theory there is no reason why it couldn't be used for any general mailing lists like Yahoo! Groups or even set of individual email address manually typed in "To:"
I'm consulting for a really large client, teaching teams how to deliver technology quickly, so I'm not working on my startup. I don't believe you can do a startup half-way.
But in my spare time I'm thinking about writing a book about how to run Agile projects in large organizations. If anybody is interested, drop me a line. My email is in my profile.
I think in general you are correct, but the devil lies in the details. The way people think about the word "agile", oddly enough, can be really inflexible. Somebody gets a CSM, reads a few Cohn books, or gets hooked on Crystal. Suddenly their definition of "agile" is exactly what they read -- no more and no less.
I read once that a programmer tried agile and gave up after arguing for two hours about what to put on the story card. This kind of BS happens a lot: it's the nature of dealing with large numbers of people who see things as binary. If you understand the _principles_ of agile, and why they work, then all the books and authors are just re-digesting common sense and peddling them as books. At that level of understanding you can work on all sorts of things: new development, existing development, non-software work, etc.
That's where my team is trying to take this organization I'm working with. The problem is that influencing large populations (thousands and tens of thousands) of technical people is far worse than herding cats. It's more like trying to get angry flaming weasels to perform circus acts by using a twinkie and a Disney song.
On some S60 code for a paying client. Improving code not originally written by me. Oh the joys of C++ and Symbian!
Lately I have been doing quite a lot of MIDP Java development for the same client, and it was quite a bit more enjoyable. With great powers come great responsibilities.
That must explain how people have already written posts complaining that it's elitist and consists of nothing more than my subjective opinions presented as if they were true.
Specifically, we just announced/launched an experimental tool for (any) Python interactive interpreter. Whenever you run into an error, you can do a quick search for how other people got past the same error.
The bug.gd site is more like a 'recycling bin' that reminds you to come back and share your workaround so it's not lost for the next guy who searches. Searching on Google works for a lot of people, but this is a lot more focused and helps the community when it's used.
Long-term we'll be integrating client-side tools into every application/language we can.
I working on mobtropolis. http://www.mobtropolis.com ,
which makes it easy to photo stream your experiences, big or small, and inspire friends or strangers to try new things.
Since then, I've added facebook integration, and not just a new look and feel, but made it more "me-centric". Just working on getting some bugs ironed out, and will be rewriting the incomplete "tour" section today.
Hey, I just downloaded your app. Nice job. I especially like how you guys seeded my tasks so that it's natural for me to start using the app instead of getting a dauntingly blank task list.
Have you guys considered a watered down free version with a lower price tag to buy the full? Like $10?
Thanks - been a lot of work! On the lower price: have thought a little but not convinced. One problem is "price-as-signal", another is that below a certain amount it just costs you too much in payment fees, marketing and support. Also, you are essentially selling the difference between the two apps for the price difference. That takes some serious thinking to get right so you don't shoot yourself in the foot...
I am also working on an MMO. It's a space game. My spin is that everything in this game will be emergent. The players can design and play anything in this game, or it will evolve through genetic algorithms. I've worked out how to do most of this. I'm still working out how to evolve NPC ship AIs, but that's not the most important.
Also, everything in this game will be scriptable. Users will be able to create programs to control anything, then license those programs to other users to make money. They will also be able to license designs. Exponential proliferation of designs and scripts is prevented by limiting designs per user to some maximum number.
Basically, I'm going to make a MMO, where the developers are out of the business of creating content. The users create the most compelling content anyhow. Oh, and the "gold farmers" -- I'm going to have them writing content for me! You can't beat market forces. Better to co-opt them.
You can think of this as Second Life in Space, with much easier implementation requirements.
I suppose you're conscious that there are a few huge cans of worms in your way. Expect QA and customer service to become more 'interesting' than in the typical MMO. Player scripting will tend to alienate nonhackers (script trading notwithstanding). Empowering players as you suggest in other comments (letting them run whole factions) reduces the pressure to socialize, which is the main driver of player retention.
Surprising as it may sound, the downsides that many MMOs have as games actually help them succeed as the glorified social networks most of them are. Boredom and disempowerment bring people together. As the cliché goes, they come for the game and stay for the community.
That said, I love your concept and I'd definitely give that game a try. I wouldn't consider playing a MMO that runs on rails and where I can't automate away my repetitive tasks. I think you'd get a smallish but interesting and loyal player base. Maybe it could reddit out from there if that's what you want; otherwise I believe it would work great as a lifestyle business.
Is it on purpose that you don't provide contact info in your profile? I'd love to exchange some thoughts with you (see address in my profile.) I worked for a few years in what is today the leading commercial space MMO, and I've been entertaining some ideas on those lines too.
Player scripting doesn't seem to alienate people in Second Life. Maybe there's something I don't know. Scripting will not be required to play. You will be able to buy scripts from weapon developers, who will finally be worthy of the name.
Faction players will need pilots. Their AI drones will never be a match for human pilots in player ships. On the other hand, pilots will need factions -- otherwise it will be hard to generate enough wealth for the very best high tech equipment.
TBH I've never played Second Life, but from what I've read it's more of a social sandbox than a direct competition. The space setting is more associated with combat and domination. It seems more likely that nonhackers will call foul on any disadvantage, real or perceived.
In my experience advocating player scripting in a MMO forum, I've observed that there is much prejudice against real life technical skills. Players seem to have less of a problem with other real life factors having an impact. Factors such as social skills or ability/willingness to spend more time on the game. I think this is because it takes technical skills to evaluate the true impact of mechanics involving technical skills, in order to come to the conclusion that they are mostly benign.
OTOH, I've also observed that some nontechnical people actually like scriptability and others can be convinced (with some difficulty; they often assume you're trying to skew the game towards your abilities for the sake of getting an edge) if you have the chance. And of course, people that can program absolutely love the feature. Unless you start with millions and a recognized IP, I think your best bet is making some underserved audience love your game rather than try to please the mainstream.
So I was playing devil's advocate here. As I said, I myself would love player scripting, and if I was to make a MMO, I'd bite the bullet. But whatever you do, I think these are factors to take into account.
Do you still have humanoid "player characters" in your game, or is your avatar simply a [set of] ships? I've always wished for a space game where each "ship" is a thriving colony all on its own, with hundreds of players with their own quarters, and an intense rivalry over who will be "captain", "navigator" and so on (unless it's some sort of military ship, but there could still be a mutiny!) If you are the ship[s], this gets a little less poignant.
Also, watching sci-fi TV shows for a long time, I've come to appreciate the visual value of "away missions"/shore-leave. It's boring staring at a starfield for hours at a time (even with cool nebular effects.) Is there a way to land on planets that have life on them (that is, not boring ol' asteroids) and perhaps stay there, living off the land for a while where no one can find you but for your radio pollution?
Sorry if this is all irrelevant to your particular implementation, but I've been planning my own similar venture for a while.
I heartily encourage you put fingers to keyboard if you can spare the time (and be sure to submit here if you do).
It strikes me as a project of such complexity and delicacy that the risks involved in publicising some of your ideas are much outweighed by the potencial gains from starting a public dialog. (though I would say that, as I'd love to read about it ;)
I find the concept of Second Life interesting, but it is difficult to use. If you could create something along the same lines of Second Life that is easier to use and eliminates the lag problem, it would definitely be something I would try.
I found Second Life after reading about Asperger's Syndrome and the communication problems associated with the syndrome that Second Life seems to break down. (No, I don't have Asperger's Syndrome; I am just an prolific reader.) What I was fascinated with is the variety of user created items and the pseudo-economy that developed within the world.
If you could create something like Second Life, it would definitely be used.
It will be like Second Life, but everything will be a Ship, an inert object in space, or an Item to be used aboard a ship. Users will be able to design Ships and Items, but only after they have built and supplied a special ship outfitted as a Research Facility and operated it for a certain period of time. Designs will be vetted at first, but this will transition to a Slashdot-like moderation system.
The galaxy will be procedurally generated. There will be millions if not billions of star systems. You will not be able to bookmark arbitrary locations in space, however. (Hyperspace is extremely chaotic and changing, so any jump to a coordinate in "Real" space will always be randomized and never precise.) You must drop an item called a "beacon" to be able to warp back to a specific location. A user will only be able to afford a certain number of beacons.
Do you intend to code a realistic physics system into this game? If so, will you be using a physics engine that has the concept of relativity built into it?
Relativity? Not right now. You will be able to fly a ship in "Newtonian" mode, but with a speed limit (due to collision implementation issues.) The ship will also fly in "inertialess" mode which will be like the way ships fly in Star Wars. Performance of ships in "inertialess" mode will be comparable to that in Newtonian, just with less flexibility.
One interesting twist to ship design, will be that the shape of your ship will actually change its stats in meaningful ways. For example, if you have a ship that uses armor for protection, then having the hull of your ship angled like an arrowhead will give you higher armor values from the forward direction. (And much weaker from the sides and rear.) Having a ship which is longer will also improve the performance of axial mount weapons. However, increased length, size and mass also have drawbacks, as it will reduce your maximum turning rate and the rate at which you can change angular momentum. Components will require some internal volume, so you won't be able to make an invulnerable Death Pencil.
The game will have elements of "space opry" and not just have hard science fiction. I want people to experience the thrill of epic space battles. There will also be exploration and trade.
Curious: how do you get the 3d models? I just wonder how much effort it is to create 3d models for a game. I am not afraid of 3d programming, but the modeling is a different kind of beast...
Right now, I'm using Wings 3d for polygonal models as in the old Elite game. I'm not an artist. I'm going to hire one or two artists to make it look good.
for the NPC ai, I recommend having the online part of the algorithm not use GAs, but just some sort of statistically learning on decision trees, GAs done right are tricky, and its worth pointing out that they're basically an optimization technique best used when the problem domain doesn't have enough structure to use other techniques which provide much stronger guarantees.
What I mean when I say that the "users make the best content anyway" is that most people just ignore game-company produced "content" or just figure out how to blow through it to get the items. What takes the place of such "content" is the real content -- players vying against players to achieve their own varied and often conflicting goals. What players have done to me and with me has excited, terrified, exhilarated, and entertained me 50X more than any game content ever has.
There are certain elements of "content" for which there will be "structured content creation." Players will be able to design new items. They will be able to play as whole factions if they want. (Using a distinct RTS style of play, controlling an entire economy of autonomous drones.) But there will be no drudgery of missions, no mining, no leveling up, no camping. Only either the action of combat, or the management of wealth generating economies.
Seems like a serious concern would be protecting the kids in this virtual world, lord knows the other "Kid" playing the game may be a 30-year old creep in his mom's basement. (Of course I worry about this when I play WoW as an adult, lol!)
I'm sure you've thought about this in depth, but my first thought was, that if you could make it a game that a child could play WITH a parent, and both coud enjoy, then that would be something I would buy for my family!
Consider movies... everyone knows the highest grossing movies are the kids movies that ALSO appeal to parents, (Shrek, Bee Movie, Shrek II, Everything Pixar, anything with the word Shrek it in...)
I'm working on a hosted load testing solution. The idea is to make it easy to find scalability issues in advance, like the signup scalability issue HN had when it was TechCrunch'd. The tests will require no hardware, be on-demand, and use an open source tool people might already be using. It should make the process ridiculously easy and scalable if I get things right.
I'm working on a web based group insurance (employee benefits) platform for the Canadian market. Right now, this market is dominated by enterprise software. Further barriers to entry basically revolve around needing to be established to be trusted.
I have the connections, and have recognized that massive inefficiencies can be capitalized on by an upstart with the right knowledge.
I'm working on a categorization/data mining project using Java JOONE for the neural network. Not as hard as it sounds (most of the time). Must finish this project today or tomorrow so I can get enough money to fuel our startup (http://www.awesomehighlighter.com) for the next couple of months.
Working on a "souce code annotating debugger" for an esoteric programming language.
So instead of painfully stepping through your code line by line, you just press M-t inside a function definition and voilà: it shows you what goes on inside the function, using some example inputs to annotate the source code in a new Emacs buffer.
My "debugger" in fact executes a whole function without stop and prints the values of all used variables and the results of nested function calls next to the code itself.
I'm not yet sure how useful this is, but it's a fun project and a good excuse to implement an interpreter for said esoterical language.
Pretty much the one thing that bothers me about working with Lisp in Emacs/Slime is the lack of what I would call a normal debugger with breakpoints, code highlighting, single-stepping and so on. Does anybody know if there's a deep reason for not having this? (I've often wondered, for example, if macros make it harder.) Or is it just that it hasn't been much of a priority?
Right now, an EEP for Unicode support in Erlang.
(EEPs are like PEPs in the Python world).
Apart from that, I've got about 10 other things on my "current stuff" list, most to do with Hypernumbers (http://hypernumbers.com), but there's a couple of Arc-related things as well.
I'm working on a member management system targeting a small niche market. Currently waiting on design revisions, my designer is coming up with some pretty unique stuff so I don't want to code too much since a lot will need to be rewritten around his designs.
4 shots of expresso with milk and foam and such, venti being the large size.
With that order though, I am still unsure if it is served hot or cold. My guess is if you don't specify, it is served hot.
Rapid application development in web applications. I'm developing an embedded domain-specific language in Haskell that basically constitutes application metadata which then gets converted to Ruby, Perl, Python, or PHP.
On my spare time I am working on a litle Android app. Financial and stock market information. Just for fun. Not going to make it it time to submit for the competition thou. I am one-two months away to have something useful.
I also want to work on something else, but as a project it would very involving, and I will need to work on it full time, and with partners. Unfortunately I am a H-1B servitor, and can't just take long breaks from my job.
Ah, and my full time job (mobile tv and multimedia, working on java and lua). Not bad, as I am actually working on advanced stuff.
Currently working on my lead generation business n2neuron.com (site is woefully incomplete and doesn't represent much of what we do, but ok because our business is conducted on the numerous content sites).
Working on my client lead-flow. Trying to break the 100 sales/day mark.
Setting up new content sites to generate more free advertising.
Right now I'm in the middle of a study group that I started here at work. We're studying Graham's book "ANSI Common Lisp" and studying inference in particular, so we're also covering some of "Paradigms of Artificial Intelligence Programming".
After that's done, I'm going to either spend some time working on the Netflix Prize (just to learn more, not in a real attempt to win) or doing some music programming in either Lisp, using the Common Music API, or Mathematica.
I rather like your site, but it seems very confusing. Like I can't find any hearts for me to click on to say that I like it. (I'm on Opera 9.5 beta.) Also, there are songs embedded everywhere, and I'm not sure why, or what all the links mean. (I don't understand their purpose.) What would also be nice would be the title and the artist within these links as well.
I do like all the music I've clicked on so far though. . .
Edit: It does have the artist and title in the embedded link. Just seemed weirdly placed. My bad there.
I'm sorry you found it confusing (hence the redesign in progress ;) ). The hearts are white until you favourite a song. I can see how this can be confusing when they are red in the instruction text.
It's not a startup, but it's something I've been thinking about doing for a while. It's a collection of articles on the topic, "Things I wish I had known before my 1st startup" You can check it out at bizglue.wordpress.com. A lot of the articles came from YC, but several came from other, random sources. It's not really a blog - more of a personal business catalog to help me in my next startup.
The online store/catalog in AWDR was non-interesting so I'm iterating a clone of Buxfer's iPhone webapp (while I teach myself Ruby-on-Rails). I have no intention of competing against Buxfer - I use their app for real expense reporting instead of my crude app - their iPhone application is a excellent example of a iPhone webapp.
Trying to write an ms exchange killer.
So many languages I want to learn, and never seem to have any time to start. (Right now on the list: Ada, Erlang, Python, Ocaml, C/C#/D/ObjC, Effiel, Smalltalk).
Was trying to do the android competition but too much junk happened in my life the past month that kept me from starting that idea.
Setting up a cash bonus plan for my team to clean up our code base. $5 for quarantined methods or blocks that just need to be deleted, $10 for deprecated blocks that need to be extracted from active code, and $25 for useful rewrites of badly designed methods. 3 days cut the fat code cash cookoff!
There have been quite a few studies that indicate that trivial reward systems tend to backfire, and have unintended side effects. These are definitely trivial rewards for good developers. Have you instead considered just implementing the measurement system, without the rewards? People can then see their success rate and have clearly defined goals for doing the work you want done. Then, perhaps take everyone out to a really good lunch, or have a catered lunch, at the end of a successful week of pruning? It's no more expensive than your reward system, and feels more like a prize to most folks--everybody says they like cold hard cash, but once basic needs are met, things like the work environment take on more importance. For example, Google pays lower salary than almost any other employer in the area, but it is the "best place to work" in the country.
When offering large enough rewards, you can convince people to do things they don't want to do. And when you ask people nicely for help, and give them clearly defined goals, you can also get them to help without a reward--they'll probably even enjoy the work more, because the human brain shoots out good chemicals when working within a community for a common cause. but, when you offer them a trivial reward, you both offend them and only grudgingly get results (with unintended consequences like code that gets written with the intention of claiming a reward later to clean it up), and morale will go down. Humans are fickle creatures, and I think you're setting yourself up for a long painful lesson in human resource management.
A centralized syslog database with rule based systems to block malevolent sorts at our firewall boundaries. Didn't like any of the solutions I found to scratch my itch.
(Ok,not technically a start up, I started up many years ago. Been there, done that, forgot to sell the company.)
Doing freelance graphic design work so I can save some money for a startup with a killer, obvious idea, but no one has done it in Jamaica yet... I'm talking disruptive stuff.
But then many great ideas in history seem obvious in retrospect.
I do visual effects and animation. Currently wrapping up a Pringles commercial. Can't show the spot but my other work is here: http://chriskelley.tv/work/
I'm working on a webapp to let people create their own games. Just finished e-mailing with a prospective cofounder (he turned me down, sigh), now going to start coding up some enhancements to collision detection.
Experimenting on a web app for people to "quickly and easily" share the stuff in their heads either with words, numbers, or pictures. Very vague at this point. Still working out some details.
Building a better dzone, mostly just to learn the Apache Wicket framework. I like the idea of dzone, but I hate the layout of the site, and I can't be the only one.
I might re-implement it in Django though, just for fun.
Customizing a Live Linux CD (for a client) writing a press release (for a client), finishing the SEO for a website (for a client) and teaching myself Ruby (FOR ME).
Today I was going to add a few features to my web app. Instead, I guess I'll be changing 20 year old programs to not update databases, but print reports. I think. Not sure.
So I thought I would just have a cup of Columbian Supreme, a donut, and a quick visit here. Makes about as much sense as anything else.
(If only this were an April Fools joke.)