Interviewee: Hadar Pedhazur
In this interview we talk with Hadar Pedhazur from Opticality Ventures. In specific, we talk about:
- The investment viability of open source companies
- Innovation around open source business models
- Customer value derived from visible source
- The potential concerns around the open source investment bandwagon
- Mixed source business models
- The future of commercial open source
Scott Swigart: Hadar, could you please take a moment to introduce yourself?
Hadar Pedhazur: Sure. I started life out as an actuarial trainee. Math was my first love, and I didn’t really know much about computers. But on the job, they forced the trainees to program for the entire firm, so they were like code slaves. I completely fell in love with computers and ended up switching to be a full-time programmer to work for other actuaries. They couldn’t believe their luck in finding someone who preferred to write programs but had actually passed actuarial exams.
From there, I ended up on Wall Street completely by accident, because I fell in love with VM/CMS from IBM as an operating system. The only thing I told the headhunters when I was looking for a job was that it had to be on VM, and a job just happened to be on Wall Street. I knew nothing about stocks, bonds, investment banks, and things like that.
I ended up having a fairly long career there–about 16 years. The last seven years at UBS, I was running a global technology group for a front-office business unit, so I didn’t report in to the back office. That gave me tremendous exposure to the business side as well. I was bringing the technology perspective but attending meetings for a global group that at one point was actually the most profitable group in UBS.
Because of that, I ended up in a position to want to make one or two personal investments in small companies when the Internet was more known by academics than by consumers–back in ’96 or early ’97. I got permission from UBS to start looking around, and during that process, UBS ended up merging with Swiss Bank Corporation, so UBS and SBC combined to become one of the largest banks in the world. That gave me an opportunity to leave, and I’d fallen in love with the idea of investing in small technology companies. Through a series of decisions along the way, I ended up opening a venture capital firm in 1998. That was called Verticality Investment Group.
I made several investments over the course of a year and a half, and then, after some disagreements with partners who I brought on, I ended up leaving that firm and starting a second venture capital firm in November of ’99 that was called Opticality Ventures. That’s what I’ve been doing for the last almost 10 years now–running Opticality Ventures.
We have one fund; it’s fully invested right now, and my only partner is my wife. At the moment, we’re down to two active portfolio companies, so we actually spend the overwhelming majority of our time working with and for Zope Corporation, which was an investment we originally made in October of 1998, so we’re coming up on our 10 year anniversary at having invested into it. All that takes you from the earliest beginning when I fell in love with computers through what I do today.
Scott: What has been your history of investing in companies that do stuff around open source?
Hadar: The percentage of companies I’ve invested in that had any involvement with open source is meaningful but on the small side, but I’m the one who actually convinced Zope Corporation (at the time they were called Digital Creations) to open source their stuff. That was back in November of 1998.
A lot of people give me credit for noticing early that good things could come from open source, from an investment perspective. I’ve also been an investor in Active State, which got sold in 2003. They were very big on the Perl side with Active Perl and other similar things, and they were very, very big users of open source at the time. So, I’ve certainly been involved with companies where the business model revolved largely around open source, although more of my investments have been on the proprietary side.
Scott: From an investor’s perspective, what makes sense about an open source model? You mentioned that you encouraged Zope to go that direction, and you’ve also been involved with companies that are proprietary, where it made sense for them to stay proprietary. What are some of the things that go into those kinds of decisions?
From an investor’s perspective, what are some of the things you look at for whether or not an open source model makes sense? Obviously, both “open source” and “closed source” are big generalizations, because there are a lot of different models within each of those, but talk about that a little bit, if you don’t mind.
Hadar: If anything, I may actually be a mirror image of what is being done at different points in time with regard to open source, and possibly even what should be done. By that I mean that, as an investor, I was fairly early in 1998 when I suggested that they open source their stuff. Really, Red Hat and VA Linux were about the only interesting things out there as companies that were open source.
There are obviously other projects that were out there forever like Python, Perl, MySQL, and things like that, but in terms of companies, there were very few. At the other end of the spectrum, right now, open source investing couldn’t be hotter. People are tripping all over themselves to get their money into any company that has some legitimate claim to being part of the open source phenomenon, and lots of the companies that I see them chasing and wanting to get their money into seem to me to be ridiculous bets from an investment point of view.
At the same time, we’ve shifted our focus at Zope Corporation dramatically. We still are heavily involved in the open source side of Zope, but our business actually has nothing to do with that. That’s what I mean when I say that I seem to be going against the herd at both ends of the exchange.
Scott: That’s what makes it a great perspective.
Hadar: When we decided to open source Zope in 1998, the thinking actually couldn’t have been simpler. I claimed zero prophet status. We had an application server, which was written in Python, and it was, in my opinion, one of the most interesting pieces of software. I was wildly impressed with it as a technology person.
But at the end of the day, Java was just starting to take off–I think in 1998, it was still immature. Clearly, it has passed that problem a long time ago. At any rate, back then, it was obvious that it was going to make it, and the interesting application servers on the market in 1998 were all written in Java.
As far as I was concerned, to run in to an enterprise customer that was our target and say, “Hey, buy our commercial proprietary application server written in some language you’ve probably never heard of. Go train your people in it. Oh yeah, you’re completely beholden to us” was a non-starter. It just didn’t make sense to me.
We also felt that no matter what we charged for the application server, we were likely to dwarf that amount of revenue in terms of how must customization we would have to add. Again, partially because people would be unlikely to have an army of Python programmers.
At the time, people were falling all over each other to keep up with their competitors trying to get on the Internet. Time to market was the really important key, which meant we knew that we were going to get services business. My thinking at the time was that if we were going to get 10 times the revenue out of any one customer from services rather than the product, let’s lower the barrier to get the damn thing in their hands to begin with.
Let’s give it away. Let’s give them the source code so they don’t feel single-vendor lock-in. Who else is going to know this application server better than us? This isn’t a case of someone coming in to support an existing open source piece of software. This is someone birthing a complete and working project overnight, with them being the only ones who can really claim expertise in it. And that worked out exactly the way we had hoped and expected. We were getting contracts hand over fist. We were doing very, very well, until the economy sort of ground to a halt.
That’s when we sat down and really thought out the problem. Other companies have proven to be successful at it, but we ended up deciding that when we grew up, we did not want to be a pure services company for the reason that the only way to grow your revenues is to grow your body count.
That means that in any roller coaster scenario, the pain is more than we would want. We did not want to be in the position of laying off 30% of our staff because we didn’t get a certain contract and then going back to rehire them six months later.
So we hunkered down and said, “What can we do with open source?” We still completely believed in open source, but what could we do, model-wise, to change that? We had tremendous fits and starts, and people could look and say, “Oh my goodness, they were all over the map,” but it was actually a strategy. We had a very strong balance sheet. We had raised a lot of money in 2000, right before the collapse of the economy, so we were never in danger, really.
What we did was to create a brand new business model. I believe we were probably first. We called it ‘visible source’. It basically has all of the characteristics of open source with the exception that you, as a customer of ours, are not permitted to distribute the source code outside of your organization.
You can use it in any way that you want. You can derive from it in any way you want, including building 50 different products from it. You can even bring outside contractors in to work on it, as long as they sign an agreement that they can’t take the code outside of the organization.
The way our CEO describes it is that you get a frosted cake, where the cake itself is fully open source. That, of course, you have no restrictions on. If you make changes in the open source part of it, feel free to go and distribute it any way you like. But that cake has a frosting on it. That frosting is proprietary, but it doesn’t lock you in. You get 100% of the source code to it. It’s not in escrow. You get it on day one. You can make any changes you want to it.
So we’re adding value, but we get to charge a license fee for that. In addition, we charge maintenance fees, to maintain that over the years as well. And, of course, we still get the benefit of having some kind of services revenue when people want customizations, ask for new features, and things like that.
We allow people to sign contributor agreements where they retain 50% ownership of any code that they enhance for the system, but they can contribute it back to the repository so they don’t have maintenance headaches in future years, if its a feature that we think is a good idea that we accept back into the repository.
It actually works almost identically to open source if it there’s a critical mass, meaning that if we had 100 large customers who were actively contributing, the world wouldn’t see it and know about it, but you’d get all the basic benefits and economies of open source.
We had fits and starts, and we actually birthed maybe five or six different products under that model. The economy was bad, and we didn’t know whether we could afford to bet on a product and market it widely. We actually didn’t market very much; we kept our costs down and just used our expertise in engineering to put a couple of different things out there.
We did OK with a handful of them, in terms of at least recouping a reasonable level of costs. We were losing money every month but keeping our losses manageable, because we had some relatively safe revenue streams.
Around that time, one of the customers for our product Zope for Media came to us and said, “Hey, listen, if you guys are willing to host this on our behalf, then that would be great. We’re having too many battles with IT. They don’t understand your software. They don’t understand Python. They don’t understand open source, and they don’t really want to maintain it. Every time there’s a problem with the server, they just point fingers and say they’re not going to fix it.” Obviously, we know how to host this stuff, and we were happy to do it.
That was the birth of our current business, and it took us four or five years to make the full turn of the aircraft carrier, but now, probably 95% of our revenue comes from hosting a whole suite of products around our Zope for Media offering. We no longer sell visible source software, with the sole exception of something called “ZRS”–Zope Replication Services.
Everything we do is basically adding value to our own software by hosting it on behalf of our users. We still get all the benefits of the open source part of the stack. Zope for Media is still the icing on the cake, so we are firmly committed to the open source part of Zope. I haven’t checked in a couple of months, but I believe it’s a true statement that we are still the largest contributors to the open source Zope.
So, we haven’t somehow changed our philosophy and stopped spending time to clean up code and put it in the repository and make it available to other people, but we don’t put the icing bits in there.
The stuff that we think differentiates our specific application, we hold on to. But anything that would benefit a wider group, we make sure to get in to the open source repository.
Scott: Well, there are a lot of things to dig into there. One point about visible source is that my conversations with people suggest that very few people actually use access to the source to solve problems or derive work from having the source.
That’s not to say that doesn’t create an opportunity that is in and of itself important, but it’s interesting what you guys did. I don’t know what you think of it, but it almost seems that a lot of the proprietary entities are slowly inching toward that type of model. Maybe not nearly as open as you guys were, but the idea of making the source code available for a variety of things they do.
They’re not going to let you go build something completely like their stuff right on top of it, but they’ll make the .NET framework code available, they’ll make certain things available to you that have a similar line. Do you think what you did is something that various proprietary organizations are trying to get to, even if not to the same degree? Maybe they’re trying to find some middle of the road where they don’t have to give away the entire farm, but they give away something that moves them toward that goal?
Hadar: I think that may be true, but I think when they’re doing it nowadays it’s usually more from a defensive posture. I used the word “software escrow” before. I would say, in this context, it’s almost like a modified software escrow, saying: “You don’t have to worry if I go out of business, because you have my source code.”
Because we weren’t marketing heavily, and the economy was really in the dumps in 2001 and 2002 when we were first launching this, we did not derive the benefit of our message. We were saying: “Well, you can go out and buy SharePoint. There’s no reason to knock what SharePoint does out of the box. But the day that you need just one more function to enhance the hundreds of things you’ve already done in that and you get stuck, forget about it! You’d better resign yourself to live within whatever boundaries they set on you.”
And the point we were trying to make is that we may actually do fewer things out of the box, but if you hit a roadblock, you have the source code to fix it, and you’ll never be locked in. Let’s use a silly example. XML RPC was hot at the time we released Zope. Let’s say tomorrow something else replaces it.
Well, it’s possible that SharePoint will never support that, because maybe Microsoft doesn’t believe in that standard, and maybe we wouldn’t either. But go into the source code, pay a consultant to come in who wrote the definitive white paper on this new protocol, let them implement 40 lines of Python code, and boom! Now, you can publish in this new protocol as well, as an example.
People that are now talking about putting some source code in people’s hands, including Microsoft, seem more defensive. I don’t believe they’re using it as a tool to try and actually give more power to the users, which is really what we were trying to do.
Scott: To a large extent, I concur with that–you can’t necessarily say that what’s happened to date has been done as an egalitarian gesture, without prompting.
On another topic, you said earlier that you see people making some ridiculous investments. What characteristics make for a ridiculous one? Is there something unique to kind of certain types of projects or the way they’re maintained or the technologies they’re trying to tackle? And what makes others a good investment?
There are a ton of them–you’ve got Xen for 500 million, and MySQL for a billion, and all kinds of other derivative things happening. As a venture capitalist, what makes the ridiculous flag go off in your head and what makes the smart move flag go off in your head?
In the dot-com bubble, everybody was trying to find an Internet company to invest in. And to me, there’s a little bit of a feeling like people are pointing at the MySQL acquisition and JBoss and Citrix and all that kind of stuff and they’re saying: “See? Open source can’t lose.”
And like you said, you’re seeing a lot of people investing in just anything open source. Is open source, to some degree, the new dot com?
Hadar: That’s exactly right. JBoss and MySQL and Xen are humongous successes, from an investor point of view, in every possible sense of the word. By the way, I actually had an early opportunity at MySQL, and by early I mean in the year 2000 or something, and I passed. I’m certainly no genius, for the record.
While you can look at those and say they are humongous successes and they are by every possible measure, fine. So were Skype and Google by any measure at all. Let’s just take the percentages of how many companies set out to be Skype and set out to be Google even within their own fields.
How many search engines set out to conquer the world and don’t exist any more? How many Voice over IP companies set out to conquer the world with a vision of having hundreds of millions or billions of users and went nowhere? The same is true of open source.
Some companies will make it, and some won’t, whether they are open source or not. The fact that someone read about a big deal happening in open source now and then chases it is just the herd mentality. Saying that open source is hot, so let me get my money into it is so generic it’s silly.
It is exactly like the dot com stuff. There was a time when companies were sold on the basis of eyeballs. In that era, when five companies in a row sold for somewhere between 20 and 50 bucks an eyeball, you are going to spend all of your time just trying to generate eyeballs. When that market collapses, then everyone who started to invest after those five companies sold is going to lose their shirt.
The analogy here is that investors think that the reason MySQL got this large valuation was because they had ‘x’ number of downloads a month. Ultimately, that’s not it–in this case, it was probably a more strategic purchase on the part of Sun. It was not just that MySQL got downloaded a lot. It is also the default back end for many important projects, including other open source projects, like WordPress.
In order to run many hot open source projects, you have to also download and install MySQL, and they’ve made that very easy. That makes MySQL a fairly interesting asset to want to control–maybe even for a billion dollars. Sun is in a different position than a normal investor to try and do something more strategic with it.
I still didn’t answer your direct question about what kind of things are interesting. I think that to the extent, 37signals is a company who doesn’t open source the ultimate edge of their stack, Backpack isn’t open source.
Scott: High Rise, Base Camp, all that stuff.
Hadar: Exactly. Under the covers, they really made their names with Ruby on Rails. In that regard, we have a totally different customer base at Zope, but we have very parallel paths in the sense that we are both giant contributors to an open source project, which brings people to look at us and to understand our brand.
We use that to add some value in a hosted manner. They use it to get as many users as possible to come to their system at a very low cost, and we use it to get a few people at a reasonably, more meaningful cost. It is a totally different sales process, but it is not a different business model, really. I think they are a very interesting company in that regard.
Other people take it a step further, and WordPress is perhaps the best example. WordPress.com is running the same software they give away for free. If you want to go and build a blog aggregation site with WordPress MU and you want to compete with them out of the gate, they’re going to give you all of the software to do it.
Not only that, but they’re going to patch it for you, and they make all kinds of other things available to you, including services like the Akismet and Blog Stats. All of this comes back and eats WordPress.com bandwidth and CPUs, and they don’t charge you unless you get to be too big. That takes the 37signals model even further, if you ask me, and I think those guys are amazing.
Scott: Do you think it’s kind of a natural evolution that if an open source project gets big enough, at least in today’s climate, it will be turned into something commercial?
Of course, the definition of commercial will vary. You’ve got Ubuntu taking Debian and turning it progressively more and more into the Linux for the desktop masses. You’ve got other efforts where they go along similar routes.
Do you think that’s natural? Sometimes I doubt that, because I look at projects like Apache, and even though you could go build a whole set of commercial endeavors on top of it, that doesn’t seem to happen. Some projects don’t seem to fall into that model, but other ones seem very likely to.
Hadar: Well, even to use your own example though, WebSphere is built on top of Apache. Part of the question there is really whether the originator of the open source project is likely to eventually decide that the best way to monetize it is to do something commercial, or whether it will it be someone else. In this case, IBM with WebSphere, Canonical with Ubuntu, rather than the original Debian guys who didn’t take it in a commercial direction.
There will be lots of exceptions, but generally, I think your point is correct. Someone will try to take those things and make them commercial. In this day and age, with the web, part of the reason for trying to birth GPL 3 was to stop an overwhelmingly gigantic loophole.
That loophole is that I could take your GPL code and offer it with a one-line enhancement, as a web service so I wasn’t distributing that code in any way. In that scenario, I wasn’t violating the license and yet I got all the benefit of all the work of the community, and I was the only one minting money.
I was following this topic for a long time. I did not read the final release, but someone I trust told me recently that they eventually had to ditch the restriction against running a commercial web service on top of it, because there was just too much flack.
So, apparently, even in GPL 3, there’s still that glaring hole. And when I say glaring hole, I mean it’s a glaring conceptual hole. I’m not so sure that all people who put out projects with GPL necessarily have a giant problem with somebody else making money off of just hosting their software, but certainly the Free Software Foundation does–conceptually, anyway.
Scott: The President of the Apache Foundation, Justin Erenkrantz, tells us that Apache’s licensing is designed to be “commercial friendly,” so you could take the Apache web server, derive from it, bundle it with your for-sale software, and you’re not required to contribute code back. Apache is basically happy with that approach.
On the other hand, like you said, with the Free Software Foundation and GPL, if you derive from it and distribute, you’re required to open source your code and contribute back. Software as a service is certainly a major loophole there.
It also seems to me that companies like Google sort of get a pass as being an open source company, but they don’t open source PayDirect or their other key intellectual property. They contribute to things like Android and some other stuff. Even Apple is almost perceived as an open source company, even though it is probably the least transparent, most secretive company out there.
On the other hand, some of the CTO’s we’ve talked to seem to understand that there’s value in their organization contributing back, because of the value that they derive from those projects. Of course, lots of organizations just use it because it’s free and feel no obligation to do anything to increase the likelihood that the open source project will survive.
Hadar: That’s extraordinarily short-sighted. The only reason not to contribute back, if you’re a large organization and making changes, is if you change something that truly could be described as your secret sauce–something about your business that you just wouldn’t want your competitor to see.
When you make some form of improvement, it’s not so much just about doing the right thing, even though that is a good enough reason to do it. It’s more about making sure that not only is there momentum, but that you end up incenting other people to share their stuff too.
If you sit around keeping on saying it’s just more painful than it’s worth, you have to realize that everyone around you is saying that as well, which means that you’re putting the burden on fewer and fewer people, and they will eventually get tired or find a new, shiny thing to go work on.
Maybe the people who are working on MySQL are not a great example, because, really, most of the code came out of the company itself, MySQL AB, but let’s just use them generically. Think about what would happen if the people who are working on that just said, “Oh, Sun owns it now. I’m going to go run to Enterprise DB.”
If you’ve made a 10-year commitment to MySQL and made changes on your own, you’ve now lost the value of the people who are not going to be there anymore.
Scott: Look at your crystal ball a little bit and tell us what you think will come from a certain evolution that we’re starting to see unfold. Open source projects like PHP and Xen start having for-profit companies wrap themselves around them, so you end up with Xend around PHP, and you end up with Xen Source around the Xen virtualization stuff.
Then those companies get acquired by massive companies, which maybe don’t have a history of open source. So, Xen Source gets acquired by Citrix. On one hand, there are certainly people in the community who take a look at that and say, “Well, maybe KVM is more interesting because it’s really a community open source kind of thing. I felt obligated to contribute to Xen when it was more of a community effort, but I don’t really feel like being a programmer for Citrix and just handing my stuff over to them.”
It hasn’t really happened yet. So far, Citrix buying Xen hasn’t really ruined it just because they were historically a proprietary company. Sun seems to understand that they can’t kill the goose that lays the golden egg with MySQL.
But as time goes on and pragmatic business guys acquire companies that happen to be wrapped around open source and not every acquisition goes well, what does your crystal ball show in the future? What happens to a Drupal when it gets wrapped by a company, it gets acquired, and things don’t go well? Does it get forked? Does it just die? Does it really depend on the project?
Hadar: I think there are two vectors that you would have to look at. Number one is what kind of mass the project had before it was bought. When you look at that vector, you say: “What’s the pain point in moving away from that for all the people who have some stake in that community?”
And so, in the case of PHP, it is very widely used on the Internet in many different ways, including in specific things like WordPress. I think the danger with a company that doesn’t have a history in open source acquiring technology with a huge mass like that is they wouldn’t have a flood of people running away until they do something wildly egregious.
In other words, even if they have benign neglect, the masses aren’t going to disappear, because too many people have their entire livelihoods tied up in it, whether it’s a business livelihood from the actual sites that are running it or because they make a living programming PHP. The danger lies in that vector because people who are new to the game will say, “Hey, I’m just going to avoid that.”
The other vector that you have to look at in this particular space, away from math, is what the company’s general intention is. Why would this company have bought them? How did they intend to monetize that?
When you have a real company that puts out financial statements every year and talks to their analysts and investors and says what their plans are, they’re going to have to justify why they bought something. If they’re trying to monetize their community in a very direct way, you might want to turn away from that particular project.
Generally, there’s a giant mass involved in any one of these things, and therefore it can’t die overnight. I think that they’re going to have the chance to prove whether they’re going to be bad actors or not. And if they’re not bad actors and it has the mass, it’s going to continue to grow.
Sean Campbell: Netscape had to mess up for a long, long, long, long time before Firefox was birthed from the ashes and grew to completely eclipse it.
Hadar: My dad got the notice that they’re no longer maintaining Netscape. He’s 82 years old and he told me he’s still not going to switch. As far as he’s concerned, that’s what he knows. He knows where all the buttons are. You’re absolutely right. It still hasn’t died. Of course it’s clearly not growing, but there are still loyal Netscape users.
Fred Wilson is one of the bigger East Coast VCs, which in itself is maybe an oxymoron. He’s based in New York and runs Union Square Ventures. Before that, he was Flatiron Partners, so he’s had a lot of big successes.
He’s an investor in lots and lots of interesting Web 2.0 companies, many of which don’t yet have obvious monetization paths. Twitter and Tumblr are his, as well as about a dozen others, all of whose names you would recognize.
He posted a blog just the other day, and then another one following it up today, talking about how innovation often stops when companies like Yahoo! buy companies like Del.icio.us, and you can see the adoption curve start dropping off. Sometimes slowly, but you see it dropping off.
It’s analogous to open source. Del.icio.us wasn’t an open source company, but in the sense that it’s a service that’s available to everybody for free, it’s similar. One of the guys inside Del.icio.us wrote back to him in a comment, and that caused Fred to have a new blog today correcting himself. He said, “Listen, our numbers haven’t dropped off the way you’re seeing. In fact, we’re growing. It just so happens that more people are accessing Del.icio.us now through the Firefox extension for Del.icio.us instead of going to the Del.icio.us website. You’re looking at Comscore numbers. You see the traffic dropping off, but it’s actually going up. It’s just coming through a different way.”
And Fred, to his credit, always has an open mind. He came back and said, “The same thing is true with Twitter. Overwhelmingly, the use of Twitter is through the Twitter API and not by going to Twitter.com and typing things in your browser.” So it was an interesting thing to see.
You might see an analogy when an open source project and a company like Citrix blend–you might see what appears to be different behavior, but maybe it’s being shifted another way, as well.
Scott: There’s a theory that popularity contests are inherently unpredictable. When you look at things like Twitter or open source projects that hit critical mass, certainly there’s traditional marketing and traditional promotion, but we run into people all the time who say, “I want to do a viral marketing campaign.” In other words, they don’t want to spend any money on marketing but want this to take off like crazy in this self-reinforcing sort of way.
I think a lot of that is luck. I think there are practically quantum fluctuations, and when something starts to gain momentum, a positive feedback effect arises, and all of a sudden everybody’s heard of Twitter.
There may have been a hundred other Twitters, and people promoted them as well or as much, but they just never made it. With as much stuff as people are doing around Web 2.0 and around open source, there are new, creative ideas every day.
Hadar: I think that what happens with these services is that when you first become aware of them, by reading an article, or through a friend, or whatever, they are so new and they are wonderfully simple, so you can grasp it immediately and start using it.
It may not be good for many different things, but you understand it intuitively, so you’re happy to pass it along to the next guy. They may say, “I think this is silly.” But it’s silly because they’ve also understood it completely and not because they gave up trying to understand it.
I think the same is true of Rails, for example. When Rails first came out, it was so overwhelmingly simple that if what you wanted to do was a basic, CRUD, form-based website, Rails made it brain-dead easy–a little bit of a hassle too, but brain-dead easy.
Eventually, though, someone decides they need a little bit more functionality, and then all of a sudden, it becomes way too complicated and it loses the elegance of its initial appeal.
Sean: You see that with development frameworks all the time. There’s a general pattern of evolution in programming languages and tools, that they inherently get more complex over time. Then there’s a new, really simple thing, which people adopt, and eventually, it becomes the next really complicated thing.
Hadar: I couldn’t agree more, and the problem is that it’s a catch-22. If you’re in charge of Rails and you keep it simple by refusing to add more features, no one will adopt it, because they’ll say it has no future. Everyone will say it’s not going anywhere, and there’s no momentum, but if I have a screwdriver that works, why do I want to innovate the screwdriver?
Sean: That point will ring true for anyone who has been in the process of disassembling a laptop and discovered that they need to run out and buy a number 0 Torx screwdriver. It’s aggravating–they’ve changed the screwdriver for no real reason, adding complexity without value.
On the other hand, things like PHP have managed that evolution fairly well. They’re still incredibly popular and incredibly simple, but they’re on version five or something.
Hadar: I was never formally in the PHP community, and so the statement I’m about to make may not be entirely fair, but I’m going to guess that’s because of all the popular products based on PHP. When millions of WordPress people upgrade and are required to upgrade from PHP 4 to PHP 5, the momentum continues and the downloads continue, and web posters around the world have to make it available in cPanel and everything else.
I keep repeating WordPress, but there are dozens of hot products out there that are PHP-based, and I’m not sure that the same would be true if it was millions of websites which were coded from scratch in PHP. You might find a lot of people saying: “You broke my site.”
My feeling is that I’m not sure the momentum derives necessarily from how cool PHP is, but from how cool the things that propelled PHP to the forefront are.
Sean: Thanks for taking the time to talk with us today. This has been a great conversation.
Hadar: Thank you.