Category: technology

Chatting and messengers

So the wife has just moved abroad and I haven’t even bothered getting international calling enabled on my mobile phone. It’s not that I’m not concerned about keeping in touch with her – it’s more to do with the plethora of options to keep in touch with her than a normal phone call.

Firstly there’s whatsapp, which I’ve used for the last two years (the trigger to join whatsapp was the limit in the number of text messages one could send per day which was introduced in 2012 as a “rumour prevention mechanism”). A large number of people on my contacts list use WhatsApp, which means that it is extremely rare that i use normal text messaging to connect to them.

And earlier today, while she was waiting for a connection at Frankfurt airport, the wife asked me to install Viber, saying it allows us to talk without any international dialing cost. I just had a brief conversation with her and the quality was extraordinary (especially given i’m on a weak BSNL broadband here and she was in a car there). Then I looked at my contacts who are on viber, and the number of my contacts who are using Viber is insanely high! Almost makes me seem foolish for not joining in so far.

And then earlier today I spoke to someone in Singapore using Skype. Call quality wasn’t that great – we dropped a couple of times – but it was still pretty good. And then there is google hangouts. And then there is apple’s facetime (perhaps the main reason the iPad fell my side when we were dividing our assets prior to the wife’s move is that I could have an Apple device with me so that we can FaceTime!).

The number of options for messaging is so large that I wonder how long the whole calling and messaging model will continue. I had shown in a recent blog post (on my public policy blog) that the number of SMSs sent per user in India peaked three years ago and has then been on a secular decline. And now there is news that the telecoms regulator in India is thinking of instituting a fee on providers such as WhatsApp and Viber because of the revenue losses they are causing to the mobile phone service providers in India (like Airtel, Vodafone, etc.).

The question therefore is what the future of telecom will look like given the large number of internet based reliable communication providers who are springing up. My prediction is that the phone call is not going to die – what sets apart a phone call from a Voice over IP connection (such as Skype or Viber) is that it is “online” (i forget the technical term for it – ok got it it’s “network switching” as opposed to “packet switching” which is how the internet works).

To explain that in English, when I talk to you over the phone (normal phone call) there is a dedicated line that goes out from me to you. Basically your telecom provider and mine and the network interchange come together so that a virtual line is drawn from me to you, and this is exclusive for us as we talk (call dropping on mobile phones happens when we try to move from one “cell” to another and get lost in between).

The internet doesn’t work that way. When I send you a “voice message” over the internet, it goes one hop at a time. There is no dedicated line from me to you. The reason we are now able to voice chat online reliably is that the bandwidth available is so much that packets usually get connected quickly enough (think of a bus network so dense that you can change buses instantly to get to your destination – it virtually simulates a “direct bus”). When the network is busy or the bandwidth clogged, however, there might be some delays (while a phone call once connected remains connected).

Given this distinction the phone call offers a level of reliability that packet switching based voice messengers can never reach. And there will always be a market for extremely high reliability. Hence the phone call is not going anywhere.

The SMS, on the other hand, is again packet switched, and a mechanism in which carriers could extract large amounts of money. The SMS will soon die a natural death – kept alive only by means of government mandated services such as two factor authentication of credit card transactions.

While the fees on carriers such as Viber might become a reality in a place like India they are unlikely to sustain as international norms become uniform. What we are likely to see instead is mobile carriers coming to terms with existence of such providers, and some interesting internet pricing plans.

Currently, to use Viber for a fair bit you need a fairly high FUP (fair usage policy) limit on your phone (carrying voice digitally takes a lot of bandwidth). Carriers might introduce some kind of a graded payment structure such that they can partly recover (through higher internet charges) the lost revenues thanks to lost call charges.

If any mobile phone operator is reading this and needs help on devising such pricing mechanisms, feel free to use my consulting services. Among other things in the past I’ve done revenue management for airline ticketing and cargo (the holy grail of revenue management) while working for Sabre – the pioneer in revenue management.

Job upgrades and downgrades, and LinkedIn

I think I’ve ranted about LinkedIn here before. I’ve talked about the pointlessness of LinkedIn recommendations (due to selection bias), the further pointlessness of skill endorsement (a desperate attempt by now-public LinkedIn to get users to interact more with each other) and the seemingly ungrammatical “say congrats” (some of these rants might have been on twitter, so not bothering to pull up links).

This post is again about the “say congrats” feature on LinkedIn. When you change your job (or, change your job title on LinkedIn), your contacts see the change on their timeline, with a helpful “say congrats on the new job” hint.

Now, the problem is that not all job changes are upgrades! Sometimes, you might get fired and change your headline from “XXX at YYY” to “ZZZ industry professional”, and LinkedIn asks your contacts to “say congrats”. Another time, you might get tired of your old job, and boldly state on your LinkedIn headline that you are looking for new opportunities (eg. “Software Engineer at XXX, looking for new opportunities”), and LinkedIn again jumps the gun and asks your friends to “say congrats”. At other times, you might make a job switch which looks eminently like a downgrade (especially for people who understand both your old and new jobs). And LinkedIn rubs it in and asks your contacts to “say congrats”.

It seems like LinkedIn needs better data scientists. And people who can make better sense of how to get their users to talk to each other and create value out of a network that is well past its fast growth phase.

Disconnected Life

The last forty eight hours were spent without internet connection, perhaps the first time I’ve spent an extended period of time at home without being connected. At first, it was incredibly peaceful, as without distractions it gave me enough time to finish off in 2 days all the 3 books (none of them very heavy, mind you, and all were “funda books”) that I’d brought home from the British Library. What was incredible was the amount of time I had in general, for everything. With the internet on, there are way too many distractions. Tweetdeck buzzes every minute. You are keen to “unbold” every mail as soon as it arrives in your inbox. Out of sheer habit, you periodically check out facebook and cricinfo. Lots of time gets wasted, no doubt.

I’d be lying, however, if I were to seay that I didn’t miss the net at all. Foremost was the need to check email, which I did though my phone periodically. I didn’t bother, however, unbolding all the stuff that was there. I only checked the mails that I thought were important, and the rest were “cleared” after I got back my connection this morning (the outage was because I’d applied for a new data plan, and the worthies at BSNL (bless them) decided I should go through some pain for having put them though the pain of changing the plan).

Then, there was some research I was trying to do yesterday, and I was looking for some data, which I wasn’t able to get since there was no internet. I went out of touch with my usual gtalk/twitter friends, but since it was only for a day I don’t really mind that. Most importantly, I missed regular updates of Ranji trophy scores, since those weren’t available anywhere else. It was too much of a hassle to be only via the phone (I don’t have 3G) for extended periods of time. There was also a lot of writing I did in the period, and all those blog posts are now sitting on my hard disk. I’ll upload them one by one with sufficient gaps so that I don’t flood you.

The worst part of no internet was the loss of the “option value” to stay connected. The best part was that it gave me a lot of time to do whatever I else really wanted to do without all the distractions the internet brings. I hope to go on an “internet break” for some time every day, switching off my modem for a few hours. Hopefully that’ll help me make better use of my time. For now, I’m glad to have the 4Mbps connection!

Yet another startup idea

This time it’s an i-phone/android app. The motivation for this is the heavy advertising in the last few days for Mapmyindia GPS, on hoardings all over Bangalore. Again, I don’t know if this has been implemented before.

So this will be built on top of Mapmyindia or any other similar GPS. When you hunt for the shortest route between point A and point B, you can give two possible choices – shortest by distance and shortest by time. The former is the default choice that all GPSs currently use. This one is an app to provide the latter.

Now, each city will be mapped out as a network of intersections. And then, for each “edge” on this graph, we use data that we’ve gathered from other users of the app in order to predict the amount of time taken to travel. Of course, the prediction model is not going to be simple, and I’m willing to partner you (via my forthcoming quant consultancy firm) in developing it. It’s going to be a fairly complex model based on time-of-day, recency of data, outlier detection (what if someone stops off for lunch in the middle of an “edge”?) and all such.

So, now you have the city mapped out (for a particular instant) both in terms of distance and in terms of time, and in cases of any traffic jams or such, my system will help you find the quickest route to your destination. Should be useful, right?

Of course, the success of this app (like a lot of other apps, I guess) depends heavily on “network effect”. The more the users of this app, the better the model I’ll have in predicting time between intersections, and save you the headache of mentally trying to optimize the route to your destination each time you set out (like I do).

I’m pretty serious about this. If you think this hasn’t been done before, we can work together to get this up!

Agoda + TripAdvisor

Ok here’s a startup idea. Basically a combination of Agoda and Tripadvisor (basically a front-end combining those two backends). I’m looking to book a hotel for a forthcoming holiday. So what I’ve been doing is to search through agoda for hotels available for those days and within my price range, and one by one searching for them on tripadvisor to see their ratings and comments and all that.

Now, the deal is this: Agoda is an excellent and reliable booking engine. However, it’s tripadvisor that has the reviews that I’d trust but it neither does bookings nor has details of availability or lowest price available. Hence I’ve to keep the two windows open which is quite frustrating and time-consuming.

For someone who’s experienced in developing web apps this is quite simple I think (since I have no experience or interest in this I’m just giving the idea away). A front end that queries agoda for available hotels and tripadvisor for ratings of these hotels and then presents both together in a nice frontend. The actual booking can be done through agoda itself (to where there can be a link).

As for revenue, I’m sure hotels will advertise on this site. Problem, though is to get the tripadvisor reviews in a way that can be extracted to this third-party website without actually going to tripadvisor. But why would tripadvisor allow this since the reviews are their intellectual property and the basis on which they make their money? But well worth a try, I think!

Google Plus – Initial thoughts

Hareesh sent me an invite for Google Plus early on Wednesday morning. Thinking it’s another stupid thing like Wave, I ignored it. But feedback from twitter revealed that the product did show some promise, so later that evening I joined it. I’ve got some 150 friends already (god knows how long it took me to get to so many with either Orkut or Facebook), though I haven’t started using it yet. Some initial thoughts:

  • I like the concept of circles, and that it’s so easy to segregate your friends. This has become a huge problem in social networking, especially after all uncles and aunties got on to facebook. So far I’ve made an attempt to classify all my contacts into disjoint circles of “friends” “family” and “acquaintances”. I also like it that circles need not be disjoint, so I can make an exception to my rule and put the wife in both “friends” and “family”
  • I like that it’s a directed graph. That you can follow the public posts of someone without them having to follow you back. I don’t know why but I simply like this. I hate putting friendship requests and waiting endlessly for responses and stuff. So this directed stuff makes a lot of sense for me.
  • I need to find out how to import my blog there. Then I can close my blog feed on facebook which is infested with uncles and aunties. On Plus, they’ll be safely tucked away in the “family” circle which won’t be able to see much.
  • I don’t like being the “cut-vertex”. I don’t like being the one guy who links two subgroups of a larger group. On a similar note, I don’t like to go out simultaneously with disjoint sets of friends (i.e. two groups that didn’t know each other previously). I feel too tense trying to make sure everyone’s comfortable and clued in on what’s happening. Similar with conversations on facebook. So yeah, I’ll probably segregate my circles further and have more cliquey groups.
  • Again, directed graph means I can peacefully put ignore to people, without appearing rude. On FB, if some uncle comments and I don’t respond, he might take offense, and I’ll be cognizant of the fact that he takes offence. And I force myself to reply. On G+, if i”m not following him, I can peacefully put well left. Like I sometimes do to @Replies to me on twitter from people who I don’t follow

So seems promising. Too early to say if it’ll make me give up both twitter and facebook. I’m sure I won’t give up twitter for sure. Let’s wait and see.

Missing BRacket

Last night then-classmate now-colleague Baada and I were having a long bitchy conversation, mostly carried over text messages (SMS). As the conversation developed and grew in intricacy, several threads developed. This is not unusual for a conversation with Baada – it usually takes on several dimensions, and it always helps having a mechanism to keep track of all the threads simultaneously.

That’s when we realized how much we miss BRacket, the local instant messaging system we had at IIMB (a version of DBabble). I might have written this before but the beauty of BRacket was that conversation was “offline”. There was no chat window, and you would reply to individual messages, like you would in email. While on one hand this allowed “offline conversation”,i..e. the conversation didn’t die if one person respond immediately like it can happen in Y!M/GTalk, the more important thing was that by having conversation history in each thread, this allowed for some serious multithreaded conversation.

While instant text messaging offers the former feature (you can reply to a message several hours later and still continue the conversation), the latter feature is lost. There’s no way to keep track of threads, and like a bad juggler you soon end up losing track of half the threads and the conversation peters out.

I don’t know if DBabble is still widely used elsewhere but it’s death knell in IIMB was sounded when Sigma (the student IT club) in its infinite wisdom allowed for a “chat mode”. Along with the conventional offline messaging system, it also gave the option of Y!M style chat windows. And having been used to Y!M, batches junior to mine started using this chat feature extensively. The immediate rewards of using it were huge – no need to hit “send” (I’ve even forgotten the keyboard shortcut for that), no need to open a new message each time it arrives, and so on.

While we held up the virtues of “old BRacket” (like i used to refuse to reply when juniors pinged me in chat mode. A notable exception being the famous “Pichai files”) there was no one to do that after we graduated. I’m told that the incoming batch of 2006 exclusively used chat mode. The two major advantages that BRacket offered over “window chat” were gone. GTalk came up sometime around then, and with its better and faster servers (the IIMB network was notoriously slow) it could easily offer as good if not better services than BRacket. It was clear then that BRacket would die.

I’m told that now no one uses BRacket. I don’t even have it installed on my last two computers. Unfortunately no other “offline-messaging” technology has quite caught on since then. And so I miss multithreaded conversation. It’s very sad, I tell you. I wonder if even DBabble is still used extensively.

It’s fascinating how some technology dies. You come up with a purported “improvement” which offers short-term gains, and catches people’s fancy. While people flock to the “improved version” in hordes, it turns out that the features that made  the original version so popular are now lost. And this new version has competition, and so the technology itself gets killed. All because of some purported “innovation”.

LinkedIn recos

LinkedIn in general is a useful site. It’s a good place to maintain an “online CV” and also track the careers of your peers and ex-peers and people you are interested in and people you are jealous of. If you are a headhunter, it is a good place to find heads to hunt, so that you can buzz them asking for their “current CTC; expected CTC; notice period” (that’s how most india-based headhunters work). It also helps you do “due diligence” (for a variety of reasons), and to even approximately figure out stuff like a person’s age, hometown, etc.

However, one thing that doesn’t make sense at all to me is the recommendations section. Point being that LinkedIn being a “formal” networking site, even a mildly negative sounding recommendation can cause much harm to a person’s career and so people don’t entertain them. Also, the formality of the site prevents one from writing cheesy recommendations – the thing that made orkut testimonials so much fun. And if you can’t be cheesy or be even mildly negative, you will be forced to write an extremely filtered recommendation.

Rhetorical question – have you ever seen a negative or even funny or even mildly unusual recommendation on LinkedIn? I haven’t, and I believe it’s for the reasons that I mentioned above. And if you think you are cool enough to write a nice recommendation for me, and that I’m cool enough to accept nice recommendations, I’m sure you and I have better places to bond than LinkedIn.

Anyway, so given that most recommendations on LinkedIn are filtered stuff, and are thus likely to be hiding much more than they reveal, isn’t it a wonder that people continue to write them, and ask for them? Isn’t it funny that “LinkedIn Experts” say that it’s an essential part of having a “good profile”? Isn’t it funny that some people will actually take these recommendations at face value?

I don’t really have an answer to this, and continue to be amazed that the market value for LinkedIn recommendations hasn’t plummetted. I must mention here that neither do I have any recommendations on LinkedIn nor have I written any. To those corporate whores who haven’t realized that LinkedIn Recommendations have no value, my sympathies.

Update

Commenting on facebook, my junior from college Shrinivas recommends http://www.endorser.org/ . Check it out for yourself. It seems like this cribbing about linkedin recommendations isn’t new. I realize I may be late, but then I’m latest.

An Inquiry Into Queue Lengths At Wedding Receptions

So last night once again I was at a wedding reception where there was a long queue for getting on stage, wishing the couple, giving gift and getting photo taken. In fact, last night, the queue literally extended to outside the hall (maybe the non-standard orientation of the hall – more breadth than length – contributed to this) – probably the first time I was seeing such a thing. Thankfully the wedding hall entrance was deep inside the building compound, else there might have been the unsavoury sight of the queue extending all the way on to the road.

This has been a problem that has been bugging me for a long time now – regarding queue length at wedding receptions. Apart from a handful, most wedding receptions that I’ve attended in the last 3-4 years have had this issue. You get to the hall and finding a long queue to get on stage, immediately go and plonk yourself at the end of it. By the time you get to the stage and do your business, you are hungry so off you go to the dining hall to probably stand in another queue. And before you know it, the reception is over and all the networking opportunity that you had been thinking of is now lost.

Udupa has a simple solution to this – introduce a token system like they have at commercial banks. Upon entry, you get a token with a number on it and you go take your seat or go around networking. And when your number gets flashed on a screen close to the stage, you go join what will be a very short queue, and you have done your business without really wasting much time. I’m told that this is the system that they had introduced at Tirupati in order to prevent time wastage at queues. However, it is doubtful if such a solution is practicable for the wedding case – people might get offended, people might get too busy to see their token number flashing, and all such.

A while back, I had raised this issue with my mom, and had casually mentioned to her about Udupa’s solution. She said that the whole problem lay with girls’ fascination for make-up nowadays, and that 99% of the problem would get solved if the reception were to start on time. This was never a problem during her time, she mentioned, when makeup was lesser and girls took less time to dress. And she also mentioned that the number of guests hasn’t gone up as significantly as one might expect.

Another solution that my mom suggested was to get the couple to stand at floor level, thus reducing the “distance” between them and the crowd, and making them more accessible. Apparently, she and my dad did that at their wedding – abandoned the stage in favour of the musicians and stood on one side at floor level, and this, she says, made crowds move faster. In fact, even at Katsa’s wedding last weekend, the couple were not at a great height off ground level, and this made them more accessible, and somehow prevented a queue from building up.

Next, we will need to look at the various processes that go into the “proceedings”. So you meet the couple. One of the couple introduces you to the spouse. You make small talk for a couple of minutes. You handover the gift. Then, you stand with the couple and wait for the photographer to make sure everything is ready, and then get your snap taken. And then put exit and head for the grub. So we need to figure out which part of this whole process needs to be reduced, or even done away with.

Gift-giving takes minimal time, so it stays. Introduction is the reason you are there at the wedding, so that also stays. Yesterday’s wedding, they took photos side-on while we were putting small talk. And that still did nothing for queue length. But still, I think that’s a good start – too much time is wasted anyways in organizing gumbals for photos. And the closer gumbals can wait for beyond grubtime.

Small talk? Is there any way that can be reduced? Two weddings recently, the couple has promised to put small talk post-reception but reception has carried on for too long making us put NED before the talk. People kept streaming in even after 10pm. Will the couple abruptly getting off stage at the closing time help? People who come later can seek out the couple wherever they are, and in the meantime they can put the small talk. And this promise means that they don’t have to put small talk when there aer 100 people waiting in the queue?

Any other bright ideas? This is a common problem. Only thing is no one party will pay you enough to come up with a brilliant solution for this – benefits of this are far too distributed. Anyways, your thoughts on this, please.