Winning but losing

Posted on May 15th, 2012

This is not one of those tech posts. There is no idea here. This is a rant post, if not philosophical!

I started out to win two things. It generally looked like winning one would make the second easier. I fooled myself into believing so. I won one thing. At least, I think I did.

The second thing is definitely not linked to the first. I still don’t understand why I presumed they were. Now am losing the second. It looks like I’ll have to lose it to keep the first victory. Wow! Did I just kill my dream?

Starting up is not easy. Killing a dream is just a few days rant. Losing tames you into accepting the world around you the way it is. I wanted a different world. I guess I ended up holding on to this one instead.

Categories: Misc, philosophy

Queue Manager

Posted on March 31st, 2012

This is inspired by the description of how starbucks manages its customers in some of its outlets as explained by a good friend of mine. The idea is to have a small handheld device that would alert the customer when his/her order is ready. The customer can then spend time in the waiting area rather than crowding the counters.

We need this in India – almost every place other than the coffee shop: clinics, banks, utility bill payment counters and so on. The basic idea is to prototype a simple hardware solution to work over an existing wireless infrastructure (or) by adding a very minimal wireless AP infrastructure. The hardware device would then be able to alert the user when his turn is due.

The more interesting addition to the idea is to implement the hardware solution over a smartphone app (i.e) an infrastructure by which the user can seamlessly register his turn in a queue and get notified over his smartphone when his turn is due. This approach has more advantages in that there is no cost of the hardware involved. There may be other network and/or implementation specific challenges to address.

Categories: electronics, mobile, technology

Grocery Shopping – Almost Online

Posted on March 29th, 2012

I came across this article in Pluggdin today. I must agree with the author because I’ve been in this situation even while shopping at a self-service store like big-bazaar. I would know one way of identifying an entity that the store staff would totally be unaware of!! Anyway, the suggestions the author makes are very relevant even if it wouldn’t solve my particular case.

Reading the article gave me another idea – another approach to ordering grocery with localization handled differently. For a large part of the non-urban population that does grocery shopping, the possibility of an internet enabled computer when they need one is pretty low. There may be desktops that they are not familiar with, or are plain un-interested in spending time with the internet. My mom for instance never surfs the net. She can use the home desktop (I know she only likes solitaire in the whole setup!). My mom does all the grocery shopping at home.

The immediate other solution was to have a mobile app to do this. Well, I doubt smart phone penetration in terms of apps etc are that big into the towns across the country. I would believe many of them have mobile phones. A good percentage might still have android devices – only to play the radio and some cine music. Or yeah, pirated movie clips So mobile app for grocery shopping might be good, but may not sell to the circle we want it to sell into!

So what are these people comfortable with anyway – apart from walking down to the nadar store and get it personally from annachi. He also gives them credit you see! All of them can talk over the phone. So can I then provide a number they can dial into and dictate their grocery list? Let us say I have a reasonably robust speech-to-text engine (ok, this is the tricky part. definitely not easy to assume you can get one out there). The dictated list gets added as a separate list that is then re-read to the user for confirmation. Once they approve, they can place the order. The grocery list is home-delivered on a cash-on-delivery basis or added to an utility bill (think airtel money, for instance).

That is not all. I could have user profiles based on phone numbers. Then the user can just dial at any time and add some thing to the list. The list keeps accumulating and then the user can confirm at the end of the month/beginning of a new month to place the order. For the app-savvy, you could supply a wunderlist type to-do list app that can automagically offer to convert certain entries to a shopping list and place order right from within there.

There are many assumptions here. Speech to text is the biggest. This can get very nasty depending how bad the voice quality of the incoming system is. I really need to update myself for the state of the art in this area. Besides, it is one thing to do speech to text in english, entirely another to do in vernacular languages with half a dozen dialects per language! One way to mitigate this is to employ a person to take the notes. A very expensive alternative to employ a person solely for this. But think of the other side – you could get college kids to think about earning their pocket money. Especially the rural folk should be very happy to make some buck on a part-time basis. And you could charge a reasonable fees for this service.

The second assumption is about the ease of acquiring a PABX type connection in non-urban areas. I would still think it should be possible to get this with some persistence at the BSNL offices.

The third assumption is that people would avail this service for a fees. My nadar annachi is a very good friend, why should I take grocery from you? Good one. I know my mom still takes grocery from that 30 year old store even though he more often than not buys the grocery outside and gives it to mom. “We are friends for so long. How do I just like that change things?” she’d say. Am sure the average grocery shopper thinks this way too. So, the only option is to avail the service at the shop of your choice. You could also mention where you want to pick the grocery from and we’d stick to that. If you had a running account, we might as well add this bill to that. Of course, you’d still owe our fees separately.

Now, do you think this will sell? With some modifications at least?

Categories: idea, mobile, technology, web

Tagged: e-commerce, grocery, ideas, mobile, online, shopping, speech-to-text, web

Car Pooling

Posted on March 26th, 2012

Car pooling is often times advised all over the place with very little takers – in India. I saw a video, quite a long time back, on either discovery channel or dw tv where car pooling was implemented almost like a call taxi service. I don’t quite remember if it was a concept video or a fully working implementation. The idea was that, if I were to travel from A to B, I would call up a certain number, identify myself, give my coordinates and express my interest in joining a car pool. Similarly all car owners who are interested in sharing the ride would register too. So when I call, if there was a car already taking that route, I would be given the identity of the car/driver and be asked to join the ride. I am assuming the driver would have some id details of myself and I would get some id details of the driver.

Even today, I’d think, there are quite a few apps out there that support/aide car pooling. I haven’t tried one myself yet. One definite reason for me not to try this is concerns of privacy, and a trust (or the lack of one) on the others that I’d have to team up with. Silly, you could say!! I was wondering how this situation could improve. For daily commute to work, there may be a slightly more trustable solution as facilitated by the company. Supposing the office allowed a sort of a social network where you get to meet all the other car users in-and-around (or) along the route you take to work, then you could personally get to know many of them to decide if you trust them. Then there is an app authorized by the company that lets a person request car pooling and a vehicle owner choose to have others join him in his way to work. Theoretically, let us assume this is workable with minor inconveniences – like there are pick-up points along major roads where you join a car pool. This way your actual residence is still kept private – if it mattered – and you could have the app validate the identity of the vehicle owner against his office identity card (smart card reader? bar code scanner?). More safer way to car pool? I think so. There is going to be a lot of effort convincing the companies to choose this one, though! Would you use this in your office?

Can we bring a similar situation onto the cab space? For instance, when you are a single person hiring a cab and assuming there are others taking the cab willing to share the ride in the same route either in part or full, can we enable car pooling with minimal detour? While minimal detour can be easily handled, how should you handle privacy? One definite way would be to have the identities of all passengers validated and recorded centrally such that it is easily accessible to authorities. Cabs are GPS tracked all the time, etc. Assuming we could convince a reasonably big cab provider to do this, would you avail this facility?

I think there is more application to this than just car pooling. There are so many other spaces where privacy and safety have been a big hurdle in technology enablement. One of the above ideas may make a small dent in that space as well. Think this through!!

Categories: idea, mobile, social, technology, tools

Tagged: cab, car pooling, enablement, GPS, ideas, identity, mobile, privacy, safety, technology

Whiteboard

Posted on January 22nd, 2012

Whiteboards are essential and are everywhere: classrooms to conference halls, todo lists and announcement boards. But very little technology has gone into these boards. Some classrooms have gotten electronic smart boards, announcement boards have become LCD displays given the rate at which the prices are falling! IIT had (and probably still has) smart boards in some presentation halls. These boards could project a document and capture what is being written on the board – sometimes annotate the parent document itself. But these things are part of any standard touch enabled gadgets, so the novelty is probably not very impressive today.

There are some reasons why smart boards are not so popular in India (or so I assume. perhaps they already are!). Cost is one big factor. If we manage to capture the essence of the smart boards but keep it really simple to reduce the cost, maybe we can improve the penetration? Anyway, I don’t have plans to compete with the smart board guys anyway.

My idea, simply, is a white board that can recognize hand-writing. It should recognize more complicated handwriting than just english alphabet: I want it to recognize greek alphabet, math, symbols, local language scripts etc. These are perhaps very essential to a class room. It need not back annotate stuff. But I would definitely want it to connect to a host PC that would just copy the recognized content into a file. Technically, hand writing recognition is just redundant here, but I want it anyway. Additionally, it would be great if the setup can record voice and have it in sync with the content that was being written on to the board. If you did this, you can immediately record lectures for archival and revisits. You can host the saved file online as notes. What’s brilliant? Well, nothing. You can do this stuff (and I know a few Profs at IIT who do) with a decent touch enabled laptop and projector. Nope, nothing brilliant here.

Can you add a small gadget to this whiteboard such that it can sync the writing to my calendar/todo list app on the cloud? Now this is useful right? And probably non-existent. Add a small protocol such that when you use a certain syntax/semantic while writing to the board, you can activate it to send an email out to a predefined address, add/sync to a predefined calendar/todo list etc. And of course, to implement this protocol you only need some handwriting recognition: either on the board or the host PC it is connected to.

I see a lot of applications for this stuff. Do you?

Categories: electronics, idea, technology

Hitchhiker’s Guide to the Earth?

Posted on January 16th, 2012

The Hitchhiker's Guide to the Galaxy (novel)

Image via Wikipedia

To a great extent, Wikipedia is the hitchhiker’s guide to the earth. I am a great fan of that book – The Hitchhiker’s Guide to the Galaxy (the book). I think it is fantastic in many aspects: the dramatic view of earth, the superb imagination of inter-galactic spaces and the humor. In fact, I’ve even felt the author taking a dig at so many things normal in our day-to-day life!

If you think about it, the book was probably the first mention of a tablet. A reader that indexed a lot of content. I wonder why the encyclopedia makers never saw the innovation. The book is essentially a socially edited encyclopedia of the galaxy. Social editing was almost entirely introduced by wikipedia in recent times. And a device to read the wikipedia was probably first introduced by Apple!

So, in not much of a good business sense, would it be a good gadget to keep that is entirely tuned to editing the Wiki? The gadget is essentially Kindle and its brilliant (what reviews say) display but doesn’t have options to actually buy books. It has the latest wikipedia and essentially stays in sync on its own and can index and display wiki pages offline. So thats an encyclopedia tablet. Add some masala to this: you know the person’s location, you know his language, maybe you have a feel for his inclinations: you could suggest links and content.

No, am not suggesting a business model for such a tablet. Just that such a dedicated wiki tablet is an excellent geek tool to keep.

– Insert – 28 Jan 2012:

Some one else figured they would like this too and has productized it as the “Wiki Reader“!

Categories: electronics, idea, mobile, technology

Tagged: Apple, Hitchhiker's Guide To The Galaxy, ideas, mobile, tablet, Wiki, Wikipedia

Network Time Protocol

Posted on January 15th, 2012

When I was a linux administrator for a small team, some years back, I was introduced to Network Time Protocol: NTP. I liked the idea and since then every time I setup a linux server/desktop, I’d try locating some NTP server to sync up to. I’ve even tried running one from my desktop to keep multiple machines in the lab sync’ed up.

After I stopped working on linux administration tasks, I seemed to have lost touch with the idea. Later after I joined my current employer I came across atomic clocks. Atomic clocks are not known in this part of the world: I doubt if we have those transmitters to sync the clocks to. I liked the idea and got my friend to pick one for me when she travelled to India. I had an idea for using the clock: nothing ground-breaking, but definitely a good way to spend a few weekends improving your design and electronic skills. Unfortunately, the weekends are yet to come. It is almost two years since I have the clock locked up in my box!

For about a year now I’ve been playing with arduino every now and then. Arduino is definitely a simple way to get your electronics jump started. With a few hours of effort, you can have many simple and moderately complex stuff going: using shields and simple C like code (sketches).

The idea is very simple: generate the required the required frequency and modulation that the atomic clock requires so that it can run seamlessly as though it were in the US. I am assuming that if you kept your power levels from the transmitter really low, you would not infringe into any proprietary/licensed bands in India. This must be worth trying out just for the understanding and getting something wireless to work. Oh yeah, I do wireless at work and I understand how crazy it is to get something to work the way you expect!!

But this has no real value in India. You anyway don’t have atomic clocks in India. Even if you get one, you need an additional something to get it to show time. This is crazy since you get simple clocks for as low as 40 bucks and you might as well spend some time syncing manually with the local radio station when they tell you the time! Even if you did decide to have an expensive clock, getting another gadget for it to work is not such an interesting thought. So, even if I do this, it would just remain in my house without many takers. I wonder if you can convince the local radio station to stream time, say NTP format, in an RDS stream that you can automatically sync to. But then that is distant and a lot of paper work to get the big guys think your idea is worth something.

This is where NTP really kicks in. Many people are working on really cool looking wall clocks and desk clocks using arduino. Search kickstarter and you’d get atleast one. Internet of things is really on the rise and am sure most arduino enthusiasts are already into this in some way or the other. Can you give your clocks an ip-address and sync them up using NTP? With arduino, the pricing should be reasonable. There are other applications to doing this, apart from NTP sync. Consider a service that you can use to securely share you calendar: am not sure about outlook and what legal hassles it will bring along, but say google calendar, or RTM and have the service push alarms onto your clock. You could work with the more beautiful wunderlist too! That way you won’t have to remember to set that alarm and you won’t have to miss the meeting because you were either sleeping or plain lost track of time doing something else!

You go a little further and add theming capabilities to your clock. Of course, that is going to make it expensive. But hey, why won’t I spend a little more if I can suit my clock to my setting and not worry about buying another if I decided to switch the setting. Theming is not the colors, but something like switching between digital and analog displays, selectively showing numbers, binary displays, scrolling displays and so on!!

You could add a mini stock ticker if you are big time into the street. Or get trending topics from twitter. Or your local news headlines if you are a news-geek. Local weather, wind speed, forecast from the met station? You could add your own sensors and line them side-by-side. Think expandability!

I guess you can dream up more things that you want the clock to do. Of course, it should display the correct time first.

Image by ToniVC via Flickr

Categories: idea, mobile, technology

Tagged: alarms, arduino, atomic clock, calendar, clock, displays, diy, ideas, internet-of-things, NTP, RDS, reminders, sync

Happy New Year

Posted on January 1st, 2012

Wish you all a very happy new year. Hopefully this year, I will keep this blog more active and updated!!

Categories: Misc

Tagged: new year, wish

Graduating from app inventor

Posted on May 8th, 2011

It turns out app inventor for android is a nice toy. It is very useful in getting you a taste android apps but can be quite limiting in what it can do for you. This is certainly not to understate all the million things you can still get done with the app inventor.

Coming from the world of web apps, I seem to need a database even to loop around values. That is one thing not very obvious with the app inventor. For two of the apps i was trying, i found this difficult to work with. So here i am jumping into developing apps the hard way

Android apps are basically java. I still have traces of java left in me though a vast majority is already lost. Anyway, thanks to eclipse I wont need them all at once

as of writing this post, i got little farther than hello world, but not far enough to yap about. so watch these pages for more yap material and possibly useful app or two!

Categories: Misc

Tagged: android, app, app inventor, database, java, mobile, web

App Inventor for Android

Posted on April 22nd, 2011

I managed to steal sometime this week to try the app inventor utility for android. Yeah, using my phone as the debug device. I just wrote two apps, one on auto messaging and another using timers. I will shortly put them up someplace for review.

What I totally loved about the experience was the ease of use. I remember fighting with symbian java for almost a week before I figured I cannot auto sms without signing the app. Signing required something short of registering a company and about two weeks in time. At that point, I moved to python which somewhat simplified it for me.

Anyway, that is just the tip. Hope to try lot more with the app inventor utility.

Categories: mobile, technology, tools

Tagged: android, apps, mobile, opinion, symbian, tools

Abishek Goda

– My personal space on the web