9 ms·

Our Odd Startup: 3½ Guys, No VC, Profitability, and Patience

97 points by sds111 9y ago

graycat 9y ago
My Odd Startup! One guy, no VC, patience, problem identified, for solution, crucial secret sauce found and scalable, production quality code written, alpha test in progress, meager burn rate, no users or revenue yet. Problem? One where the first good solution should be a "must have" for nearly everyone on the Internet. Solution? An excellent solution now with a high technological barrier to entry and later with some strong network effects. Solution appears to the users as just a simple Web site that looks good on everything with a Web browser up to date as of maybe 10 years ago and from just 24,000 programming language statements. Platform? Microsoft's .NET Framework 4.0 with ASP.NET, ADO.NET, Visual Basic .NET, and a little use of C via platform invoke. Potential? If people like the solution, then in line to be the first company worth $1T.
- radicalbyte 9y ago
  Why on earth would you use VB.Net for something new?
  - quincunx 9y ago
    To just get it done? To win dirty? To just ship it? Why would a carpenter have to learn masonry to build a house? Because it makes for a better house? He might never make the house, better to build it out of wood.
  - mathgeek 9y ago
    Because you use the tool you know best.
    - graycat 9y ago
      I learned VB.NET just for this project. I had to select: Windows v. Linux. On Windows, VB.NET v. C#. I selected VB.NET. Easy enough decisions.
      - mathgeek 9y ago
        Oh... that's unusual to hear, but best of luck!
  - graycat 9y ago
    Why not? From VB.NET, can call the original .NET Framework, e.g., 4.0, 4.5, whatever, including ASP.NET and ADO.NET. Side issue: Just what the .NET Core is and/or its pros/cons I don't know. Microsoft's documentation seems to assume I already know what the Core is and why I'd want it. Not the first time Microsoft didn't do well explaining their work. Microsoft: I don't have even as much as a weak little hollow hint of a tiny clue what the heck the .NET Core is or why I would or would not want it; the times I tried to find out were all self-inflicted tooth drilling; I likely don't need the Core or have to care about it so gave up trying to find out. Next, on Windows, AFAIK, the main alternative to VB.NET would be just C#. My understanding is that VB.NET and C# differ mostly only in their particular flavors of syntactic sugar and that there is a translator from one to the other, maybe from VB.NET to C#. The flavor of C# was borrowed from C, and I don't like it. Back to K&R, the C syntax was deliberately absurdly sparse and idiosyncratic. I wrote some C code; I mowed grass in 103 F; I shoveled snow in 0 F; I didn't like any of those three. The VB.NET syntax is closer to traditional for Pascal, PL/I, Fortran, and Algol. Compared with C#, IMHO VB.NET is easier to read, write, teach, and learn and less prone to errors from obscure syntax. With VB.NET I get (A) objects, (B) strings, (C) memory management ("Look, Ma, no memory leaks!"), (D) good enough exceptional condition handling, (E) platform invoke for calling, say, C, (F) fast compiling with good error messages and so far no bugs encountered, (G) reasonably good documentation, (H) ability to have subroutines and functions and some scope of names functionality, (I) the usual Do While and If-Then-Else, (J) the math functions and some good time and date handling functions,(K) easy comment syntax, .... Looks fine to me. I'd like more on scope of names, e.g., as in PL/I. Maybe a pre-processor, e.g., again as in PL/I, would be nice. At times longer precision arithmetic would be nice; I don't know what the latest x86 has in hardware. And I can think of some things new to programming languages that would be nice. But I can't complain: VB.NET has let me write my 24,000 statements in 100,000 lines of typing (just from my favorite editor KEdit and a lot of macros but no use of an IDE). So, with the 100,000 lines for 24,000 statements, there is a lot of internal documentation. I had no trouble debugging the code. I wrote no prototype code, and my code doesn't need refactoring. The code looks fine for quite significant production. For ASP.NET and IIS, maybe I could gripe: First, I didn't like their approaches to Web user session state. So, I wrote my own Web Session State Server, just TCP/IP socket communications with de/serialization, and two instances of their main collection class -- darned fast and easy enough to scale just via sharding, even without Web user session affinity. For the input queue, I just used the TCP/IP FIFO; the queue depth permitted should be deep enough. For a Web site log server, again I don't like what Microsoft has so, as soon as I rebuild the main boot partition on my development computer from a disaster, I'll take the Web Session State Store, rip out the use of the collection class, put in a file write statement, and, presto, bingo, get a Web site log server. In total the work had a lot of mud wrestling and fighting unexpected problems. But the work that was uniquely mine and at all predictable was all fast, fun, and easy.
  - rsynnott 9y ago
    I didn’t know it was even a thing; haven’t heard anything about it in around ten years.
    - graycat 9y ago
      Visual Basic .NET is not much like the Visual Basic of 10 years ago. As far as I can tell, Visual Basic .NET is now essentially as powerful as C# or any other Microsoft language on Windows. Supposedly C# has lambda expressions. Okay. If I need those, maybe I'll write some C#. But have VB.NET call or be called by C# is relatively easy. So, I could have a function in C# called by my main code in VB.NET. I wrote enough C to understand it fairly well, at least at the level in K&R. I've looked at some C# code, and it seems to read easily enough. Maybe I could be writing C# in a few days. If I want to convert to C#, I understand that there is a translator to do that for me. I do like the more traditional syntax of VB.NET instead of the deliberately "idiosyncratic" syntax of C borrowed by C#. So, I see no biggie reason not to use the syntax I prefer in VB.NET. I learned programming long before C++ or even C. I skipped C++ -- never wrote a line of it. My first code with objects was essentially VB.NET. The objects are easy enough, close enough to what I did with structures in PL/I. Now with .NET, VB.NET, and C#, there's reduced reason to write C++ on Windows. There are plenty of serious criticisms of C++ for important production software. Yup, it's possible to write good code for big projects in C++, but just from memory management I'll let VB.NET handle that for me.
      - ShabbosGoy 9y ago
        Good luck with your project! This was a fun read. Don't let the haters keep you down, just remember Zuckerberg shipped Facebook in PHP.
        graycat 9y ago
        Yes, supposedly the first version of Google was something equally weak! Apparently since then Google has written a huge volume of some really good code. Good for Google.
- jacquesm 9y ago
  If it was your intention to cause me a couple of sleepless nights consider your mission accomplished.
  - rboyd 9y ago
    how so? just from his posts?
- sk5t 9y ago
  If your work product has a real barrier to entry and strong value like you say, I wish you'd stop coming here to 1) play coy / not describe the product, 2) flog dissatisfaction with VC behavior, 3) throw out outrageous $1T numbers, and finally 4) rile people up with arguments about VB.NET. Look, it's nice that you got some use out of Microsoft's frankly very irritating old developer tools. But all we're doing here is reorganizing bits, any language will do that job, although it is harder to enlist others in helping if you have settled on awkward means to get there. At some point you will have to "put up or shut up" about the thing you believe you're building.
  - graycat 9y ago
    > 1) play coy / not describe the product, "Play coy" has nothing to do with it. Holding off the details of the announcement until my Web site is ready to go live and please lots of people is important and the reason for not saying more now. > 2) flog dissatisfaction with VC behavior One lesson can learn about VC behavior from my posts is that so far, from all I've written that you don't like, no VC likes to hear at all, not a tiny bit. So, a lesson for every entrepreneur on HN is that nothing I have written in any way covers what is necessary for VC interest. Why is this important? VC Web sites commonly emphasize new, advanced technology, big markets, swinging for the fences, ..., etc. Well, my posts promise all of that. But no VCs respond. The same things sent to VCs via e-mail have the same results -- dead. The same things sent to VCs via Twitter -- same thing. So, from what I am writing here, some thing necessary for VC interest is missing. IMHO: Essentially ALL the information technology VCs are interested in is traction significant and growing rapidly. Essentially everything without that traction goes in the trash. VCs don't want to say this, but entrepreneurs need to know it or do what I did -- jerk the chains of nearly every VC on both US coasts with pitches of every style. The responses were essentially all zero; essentially the only thing missing was traction. That VCs want traction is, thus, a fairly solid conclusion and really important for entrepreneurs to know about VCs. Why are VCs this way? IMHO, the LPs got together and wrote out some severe criteria for VC investments. The criteria are much like in private equity or commercial banking: Only make a financial investment in a financial asset. Technology, algorithms, running code, secret sauce, etc. are all just irrelevant. For VC, the LPs are willing to bend a little and call traction significant and growing rapidly a suitable surrogate for the usual financial assets. Entrepreneurs need to know this. But for a solo founder with the only capex needed a PC as a first server, the LPs rules are too severe. But some of the VCs and LPs are making money now. Here is a secret -- don't tell anyone! If I am successful, then I will have a long track record here on HN saying to the VCs "I told you so." or "You ignored me until I had revenue enough to ignore you." Or I'll be able to point to my posts here as not sufficient for any VC interest. Maybe then, finally, the VCs will be clear on what their criteria are, in one word, traction. > 3) throw out outrageous $1T numbers, There is nothing "outrageous" there. Instead it is just simple arithmetic. If you are willing to entertain that there is a problem such that nearly everyone on the Web will find the first good solution a "must have" (e.g., like horse owners found the Model T), then consider the 1-2 billion users in the more developed countries, that is, where can get some good ad revenue. Assume that those people will go for the solution ballpark 3 times a week and each time see 50 ordinary ads (the user interface is interactive and iterative which is crucial for the core applied math but also good for showing users a lot of ads with, it turns out, quite good ad targeting). Assume the relatively good ad targeting. Then compare with the number of unique users, the daily user traffic, page views, ad impressions, revenue, and market cap of Google and Facebook. Done. Nothing at all outrageous. Or get a lot of eyeball time each week from nearly everyone in the more developed countries and, presto, bingo, can monetize to $1T market cap. Sure, a Web site for selling winter sweaters for puppy dogs on line might be an okay business, but, as a first step in picking a business idea or project, pick a problem that nearly everyone has, not just people who have a dog, cherish it, and want to walk it in cold weather. I tried to do that; dirt simple tactic. The problem? One I had and still have, daily. I believe that essentially everyone on the Internet has the same problem. So far there is no good solution. Why no solution? Can't get a good solution with only what is common on Sand Hill Road or in the Stanford, CMU, or MIT CS departments now. My applied math is nothing like what the AI/ML people are doing. I have nothing neural. I don't claim that my code is intelligent. No one screaming for a solution? Well, before the Model T, horse owners were not screaming for the Model T, either. Why not? They could not imagine something both better and at all realistic. Essentially every horse owner had the problem, and the Model T was the first good solution. Same for my work: Internet users have the problem but can't imagine something both better and at all realistic. Besides a big theme in information technology is moving ahead! E.g., in the last few months we have seen 14 TB SSDs, 16 thread consumer processors, and, from my ISP, download data rates of 100 Mbps up from about 10 Mbps over the interval, with the prices falling quickly. Main memory is going for ballpark $10/GB, one DIMM at a time, retail. Back there, 300 MB of disk storage weighed 750 pounds and cost $40,000. Big changes. Well, I have in mind another big change but will be not nearly the first. In lots of other ways, we have seen big changes. E.g., Amazon: They were an on-line book and record shop, but consider them now. How'd they do that? Well when Bezos started, no way was the Internet data rate high enough to permit him to send so many pictures of the products. Now the pictures are so good that shopping in brick and mortar sources does not have better product information, usually much less good. Big changes are possible. But, I can assure you, essentially no VC in the country sees my $1T and is interested enough for a phone call, an e-mail, or even a Tweet. E.g., USV just got a new investing partner who posted > If you’re a founder or potential founder building something transformational, I’m so eager to meet you. I’m on a new adventure and I’m ready to dive into yours, too. You can reach me at Rebecca@USV.com. So, I wrote her. Hear back? Nope! And I'm not holding my breath waiting! So, why do VCs post such garbage? Curious. Conjecture: They want to maintain the image that they invest in things that are "transformational" without mentioning traction. So, it's an initial tactic for getting "deal flow" and then deal negotiation. > 4) rile people up with arguments about VB.NET. What's to be riled up about? As a platform, I picked Windows instead of Linux -- lots of people and serious organizations have done that. On Windows, a key to nearly any software development of production software is the .NET Framework. Okay. Well, Visual Basic .NET can make apparently full use of the .NET Framework. Indeed, Microsoft MSDN Web pages documenting the .NET Framework, and I have about 5000 of them, are all or nearly all just the same for VB.NET as anything else, including C#. So, what's wrong with writing in Visual Basic .NET? I see nothing wrong. What is the much better alternative? I see no significantly better alternative. > Look, it's nice that you got some use out of Microsoft's frankly very irritating old developer tools. What? What "old developer tools"? I wrote in Visual Basic .NET which is an essentially fully up to date language on Windows and very different from the old Visual Basic. So, what are you concerned about?
- ci5er 9y ago
  Good luck. You may be a math genius, but you don't have a marketing or venture-formation bone in your body. (I do applaud you're choice not to take on VC money, though...)
  - graycat 9y ago
    You have me suspected, accused, tried, convicted, and sentenced all in absentia and with no good evidence for doing NOTHING wrong.
    - ci5er 9y ago
      You are right. I was wrong. I am sorry.
- calgaryeng 9y ago
  Overconfident comes to mind....
  - graycat 9y ago
    You have me suspected, accused, tried, convicted, and sentenced all in absentia and with no good evidence and while doing NOTHING wrong. No: Just a simple observation, e.g., much like horse owners needed something better, e.g., Ford's Model T which then became a "must have" for nearly every horse owner in the more developed countries. Essentially every user of the Internet has this problem everyday; currently there is no good solution on the Internet. The first good solution will be a "must have" for nearly every user on the Internet in the world. And my solution is not just good but excellent, better from being more powerful than possible with the techniques common on Sand Hill Road or in the computer science departments of Stanford, CMU, or MIT. My core technology is not computer science, AI, or ML, and is original with me. Current work in AI/ML can't get the solutions my work delivers. I'm able to do such original work; that's what my Ph.D. says and why I got it. The estimate of $1T market capitalization is back of the envelope, first-cut, ballpark but with some reasonably solid evidence: The main evidence is just the prospect of getting on average, say, 30 minutes of eyeball time three times a week from nearly every Internet user in the more developed countries (where can get good ad rates). Comparing with other successful ad supported Web sites, the $1T is reasonable enough. But for being speculative, it might be that, like cars where the Model T was not the end and not even the beginning of the end but just the start of the beginning, the usage and ad revenue might make the effort worth several times $1T. And there is a chance that the ad targeting can be come a big, new thing beyond just the ordinary views of ads for several times more. But that's speculative. Sure, cars are an example of how just improving on horses can become really big, e.g., create the suburbs. So are phones, PCs, dial up modems, AOL, CompuServe, and Prodigy, radio that became TV, computers, transistor based computers, microprocessors, and 14 nm line widths, aluminum, and more. Amazon is a big change, not just in retailing but even in logistics and the economy. So, with my project, it could become much bigger than the $1T. E.g., likely the current ad business would have difficulty supporting much over the $1T, but if go beyond ads, say, as Amazon went beyond just retailing and cars went beyond just replacing horses, and airplanes beyond what the Wright Brothers did, and satellites well beyond Sputnik, then could replace ads with something much better and get still more. But that's the speculative part. The $1T is reasonable enough, for a ballpark, first-cut, back of the envelope estimate. But, my applied math has some generality for taking in common data and putting out valuable information and could be applied to and be the core of something bigger than my current goal of the $1T. Then a solution, based on some new data and some rock solid applied math, complete with theorems and proofs (I hold a good Ph.D. in such work) based on some advanced pure/applied math prerequisites. The math is nicely typed into a document via Knuth's TeX. No, not LaTeX, just Knuth's original, long essentially totally bug free TeX. It's a nice applied math document. Am I going to publish that secret sauce? Heck no! A new, innovative, very different, but simple and easy to use, UI/UX that is interactive, iterative, somewhat game like, and for some users addictive and that helps generate the new data. The UI/UX needs just a Web browser up to date as of 10+ years ago. E.g., ASP.NET automatically wrote a little JavaScript for me, but I never wrote even a single line of it. That ASP.NET is the only JavaScript for the site, and even that is optional (I suspect it has to do with cursor positioning). The Web pages, then, are just dirt simple, send for <= 400,000 bits per page, with the JavaScript so load very quickly. There's no use of Ajax. The fonts are large; the colors easy to see (about 25% of men are partially red-green color blind but will be able to read my Web pages just fine); the backgrounds are all just white; the contrast is high. The pages are just dirt simple with just a few of the simple, standard HTML controls nearly everyone on the Internet understands; there are no icons, roll-overs, pull-downs, pop-ups, overlays, etc.; the screens just load, BAM, and don't jump around during loading. The pages will look fine on everything from a smartphone to a high end workstation. Production quality code in solid Windows Visual Basic .NET (no, NOT the old, simplistic, weak, unsupported Visual Basic 6.0), essentially the most powerful language for the .NET Framework on Windows, e.g., essentially just a different flavor of syntactic sugar compared with C#, fine, about as good as it gets these days for production quality software. The code is 24,000 programming language statements in 100,000 lines of code; so, right, is profusely documented in the code but also otherwise. Code running as intended with no known bugs and in alpha test. Code intended for first production through a quite significant company, say, $10 million a month in revenue with nearly nothing in opex. So, no prototype code. No minimum viable product. No stuck together with chewing gum and bailing wire. No refactoring needed. Instead, solid, production code. My first Web site, first use of SQL Server, first use of Visual Basic .NET, first use of .NET, ASP.NET, ADO.NET, my first UI design, my first Web page design, all production quality work from the beginning with no intermediate steps, prototypes, redesigns, refactorings, etc. I typed in the code myself using just my favorite text editor KEdit. I made no use of an integrated development environment (IDE). I had no difficult debugging. Timings show that the code is shockingly fast. The data base has only clustered keys, no foreign keys, and no joins. SQL Server is being used as not much more than just a B-tree data base system. In production, SQL Server is being lightly used -- the main data manipulations are elsewhere in my code. There is an off-line, batch program that reads the SQL Server data and does some initial processing later used on-line. It's a good project, a darned good project. There's just good planning, applied math, software development, etc. and nothing over confident. I didn't get into such work yesterday or even recently. I've been successful doing applied math and computing for US national security and business for decades. I know very well just what the heck I'm doing. It all worked just as I originally intended. There were some problems but not specific to my project. All the work unique to my project was fast, fun, and easy for me. Oh, by the way, I am 100% completely, totally, absolutely, permanently unemployable for any job above stocking shelves at Wal-Mart and likely even that. I'm a native born US citizen, never arrested, in good health, never filed an insurance claim of any kind, never sued or been sued, never convicted of any crime worst than a minor traffic violation, have not had a traffic violation in 10+ years, never used illegal drugs or used legal drugs illegally, etc. Yup, permanently unemployable. For a solo founder to build a company worth $1T is exceptional; same for $100 billion, $10 billion, $1 billion (the VCs hope for), $100 million. So, one might suspect that in some ways such a person and project would be a little different from what is quite common. So be it. A good thing about the Web is that the users don't know anything about the owner of the Web site; they might be a dog, wearing blue jeans.
a_imho 9y ago
I like the takeaways, but the no VC seems to be quite an exaggeration. They invested temporary resources (design, frontend) and a full-time salary (mine) into Snipcart, our developer-first e-commerce platform.
- jacquesm 9y ago
  That's not venture capital, that's a bit of boost at the front, and most of it not in money. They didn't say they self-financed all of it, just that they did not take on any venture capital.
  - weego 9y ago
    OK so it's not VC in the narrowest of definitions, but there was capital investment by a third party so there are some language gymnastics involved to overstate the independent nature of the company.
    - jacquesm 9y ago
      There is nothing narrow about that definition, rather, you are trying to stretch what it means to take on venture capital to things it was never intended to cover. The language gymnastics are entirely on your part.
    - drharby 9y ago
      Moving goalposts?
  - a_imho 9y ago
    Semantics, there is similar argument there without the exponential growth they are closer to a (micro) saas than a startup (discalimer, I've no idea how big Snipcart is) and their return is ~on par with the investment.
- toyg 9y ago
  Looks like Spektrum was basically acting as an incubator that spent some money on them. Your criticism might be valid depending on whether the incubator got some ownership rights (in which case it was acting as VC, albeit on a small scale) or not (in which case it was just an angel investor).
wordpressdev 9y ago
Being a solopreneur, I am a one man startup.
onion2k 9y ago
Sure, there are now dozens of e-commerce solutions similar to Snipcart. Yet there is 0 business like ours. It sounds like Snipcart are doing something very similar to Moltin (https://moltin.com https://moltin.com), which is a YC company.
- tehwebguy 9y ago
  My implementation of this was rejected in 2013. I should have put more work into my YC app + video!
ojr 9y ago
So it is 4 guys not 3 1/2 just because you don't work full time doesnt me youre 1/2 a person?
- Sylos 9y ago
  Right, but then why would you click on that headline?
  - ojr 9y ago
    I clicked to figure out why a person was described as a half person but I didn't spend much time on the page and didn't figure it out, do you want to know what I ate for lunch too?
- gggdvnkhmbgjvbn 9y ago
  It was him plus the cast of two and a half men.
5_minutes 9y ago
Headline might be over the top but the essence here is a nice vibe.
zer0livemagic 9y ago
I'd agree that the fascination with VC money is not helpful. There are certainly verticals where you can only survive with outside investments (biotech comes to mind) or where scaling quickly is important, but no amount of money will save you if you don't have product-market fit.
- anovikov 9y ago
  Well, a good half of the ecosystem is tailored to eat away investor money and nothing else.

97 points by sds111 9y ago

My Odd Startup! One guy, no VC, patience, problem identified, for solution, crucial secret sauce found and scalable, production quality code written, alpha test in progress, meager burn rate, no users or revenue yet. Problem? One where the first good solution should be a "must have" for nearly everyone on the Internet. Solution? An excellent solution now with a high technological barrier to entry and later with some strong network effects. Solution appears to the users as just a simple Web site that looks good on everything with a Web browser up to date as of maybe 10 years ago and from just 24,000 programming language statements. Platform? Microsoft's .NET Framework 4.0 with ASP.NET, ADO.NET, Visual Basic .NET, and a little use of C via platform invoke. Potential? If people like the solution, then in line to be the first company worth $1T.

radicalbyte 9y ago
Why on earth would you use VB.Net for something new?
- quincunx 9y ago
  To just get it done? To win dirty? To just ship it? Why would a carpenter have to learn masonry to build a house? Because it makes for a better house? He might never make the house, better to build it out of wood.
- mathgeek 9y ago
  Because you use the tool you know best.
  - graycat 9y ago
    I learned VB.NET just for this project. I had to select: Windows v. Linux. On Windows, VB.NET v. C#. I selected VB.NET. Easy enough decisions.
    - mathgeek 9y ago
      Oh... that's unusual to hear, but best of luck!
- graycat 9y ago
  Why not? From VB.NET, can call the original .NET Framework, e.g., 4.0, 4.5, whatever, including ASP.NET and ADO.NET. Side issue: Just what the .NET Core is and/or its pros/cons I don't know. Microsoft's documentation seems to assume I already know what the Core is and why I'd want it. Not the first time Microsoft didn't do well explaining their work. Microsoft: I don't have even as much as a weak little hollow hint of a tiny clue what the heck the .NET Core is or why I would or would not want it; the times I tried to find out were all self-inflicted tooth drilling; I likely don't need the Core or have to care about it so gave up trying to find out. Next, on Windows, AFAIK, the main alternative to VB.NET would be just C#. My understanding is that VB.NET and C# differ mostly only in their particular flavors of syntactic sugar and that there is a translator from one to the other, maybe from VB.NET to C#. The flavor of C# was borrowed from C, and I don't like it. Back to K&R, the C syntax was deliberately absurdly sparse and idiosyncratic. I wrote some C code; I mowed grass in 103 F; I shoveled snow in 0 F; I didn't like any of those three. The VB.NET syntax is closer to traditional for Pascal, PL/I, Fortran, and Algol. Compared with C#, IMHO VB.NET is easier to read, write, teach, and learn and less prone to errors from obscure syntax. With VB.NET I get (A) objects, (B) strings, (C) memory management ("Look, Ma, no memory leaks!"), (D) good enough exceptional condition handling, (E) platform invoke for calling, say, C, (F) fast compiling with good error messages and so far no bugs encountered, (G) reasonably good documentation, (H) ability to have subroutines and functions and some scope of names functionality, (I) the usual Do While and If-Then-Else, (J) the math functions and some good time and date handling functions,(K) easy comment syntax, .... Looks fine to me. I'd like more on scope of names, e.g., as in PL/I. Maybe a pre-processor, e.g., again as in PL/I, would be nice. At times longer precision arithmetic would be nice; I don't know what the latest x86 has in hardware. And I can think of some things new to programming languages that would be nice. But I can't complain: VB.NET has let me write my 24,000 statements in 100,000 lines of typing (just from my favorite editor KEdit and a lot of macros but no use of an IDE). So, with the 100,000 lines for 24,000 statements, there is a lot of internal documentation. I had no trouble debugging the code. I wrote no prototype code, and my code doesn't need refactoring. The code looks fine for quite significant production. For ASP.NET and IIS, maybe I could gripe: First, I didn't like their approaches to Web user session state. So, I wrote my own Web Session State Server, just TCP/IP socket communications with de/serialization, and two instances of their main collection class -- darned fast and easy enough to scale just via sharding, even without Web user session affinity. For the input queue, I just used the TCP/IP FIFO; the queue depth permitted should be deep enough. For a Web site log server, again I don't like what Microsoft has so, as soon as I rebuild the main boot partition on my development computer from a disaster, I'll take the Web Session State Store, rip out the use of the collection class, put in a file write statement, and, presto, bingo, get a Web site log server. In total the work had a lot of mud wrestling and fighting unexpected problems. But the work that was uniquely mine and at all predictable was all fast, fun, and easy.
- rsynnott 9y ago
  I didn’t know it was even a thing; haven’t heard anything about it in around ten years.
  - graycat 9y ago
    Visual Basic .NET is not much like the Visual Basic of 10 years ago. As far as I can tell, Visual Basic .NET is now essentially as powerful as C# or any other Microsoft language on Windows. Supposedly C# has lambda expressions. Okay. If I need those, maybe I'll write some C#. But have VB.NET call or be called by C# is relatively easy. So, I could have a function in C# called by my main code in VB.NET. I wrote enough C to understand it fairly well, at least at the level in K&R. I've looked at some C# code, and it seems to read easily enough. Maybe I could be writing C# in a few days. If I want to convert to C#, I understand that there is a translator to do that for me. I do like the more traditional syntax of VB.NET instead of the deliberately "idiosyncratic" syntax of C borrowed by C#. So, I see no biggie reason not to use the syntax I prefer in VB.NET. I learned programming long before C++ or even C. I skipped C++ -- never wrote a line of it. My first code with objects was essentially VB.NET. The objects are easy enough, close enough to what I did with structures in PL/I. Now with .NET, VB.NET, and C#, there's reduced reason to write C++ on Windows. There are plenty of serious criticisms of C++ for important production software. Yup, it's possible to write good code for big projects in C++, but just from memory management I'll let VB.NET handle that for me.
    - ShabbosGoy 9y ago
      Good luck with your project! This was a fun read. Don't let the haters keep you down, just remember Zuckerberg shipped Facebook in PHP.
      - graycat 9y ago
        Yes, supposedly the first version of Google was something equally weak! Apparently since then Google has written a huge volume of some really good code. Good for Google.
jacquesm 9y ago
If it was your intention to cause me a couple of sleepless nights consider your mission accomplished.
- rboyd 9y ago
  how so? just from his posts?
sk5t 9y ago
If your work product has a real barrier to entry and strong value like you say, I wish you'd stop coming here to 1) play coy / not describe the product, 2) flog dissatisfaction with VC behavior, 3) throw out outrageous $1T numbers, and finally 4) rile people up with arguments about VB.NET. Look, it's nice that you got some use out of Microsoft's frankly very irritating old developer tools. But all we're doing here is reorganizing bits, any language will do that job, although it is harder to enlist others in helping if you have settled on awkward means to get there. At some point you will have to "put up or shut up" about the thing you believe you're building.
- graycat 9y ago
  > 1) play coy / not describe the product, "Play coy" has nothing to do with it. Holding off the details of the announcement until my Web site is ready to go live and please lots of people is important and the reason for not saying more now. > 2) flog dissatisfaction with VC behavior One lesson can learn about VC behavior from my posts is that so far, from all I've written that you don't like, no VC likes to hear at all, not a tiny bit. So, a lesson for every entrepreneur on HN is that nothing I have written in any way covers what is necessary for VC interest. Why is this important? VC Web sites commonly emphasize new, advanced technology, big markets, swinging for the fences, ..., etc. Well, my posts promise all of that. But no VCs respond. The same things sent to VCs via e-mail have the same results -- dead. The same things sent to VCs via Twitter -- same thing. So, from what I am writing here, some thing necessary for VC interest is missing. IMHO: Essentially ALL the information technology VCs are interested in is traction significant and growing rapidly. Essentially everything without that traction goes in the trash. VCs don't want to say this, but entrepreneurs need to know it or do what I did -- jerk the chains of nearly every VC on both US coasts with pitches of every style. The responses were essentially all zero; essentially the only thing missing was traction. That VCs want traction is, thus, a fairly solid conclusion and really important for entrepreneurs to know about VCs. Why are VCs this way? IMHO, the LPs got together and wrote out some severe criteria for VC investments. The criteria are much like in private equity or commercial banking: Only make a financial investment in a financial asset. Technology, algorithms, running code, secret sauce, etc. are all just irrelevant. For VC, the LPs are willing to bend a little and call traction significant and growing rapidly a suitable surrogate for the usual financial assets. Entrepreneurs need to know this. But for a solo founder with the only capex needed a PC as a first server, the LPs rules are too severe. But some of the VCs and LPs are making money now. Here is a secret -- don't tell anyone! If I am successful, then I will have a long track record here on HN saying to the VCs "I told you so." or "You ignored me until I had revenue enough to ignore you." Or I'll be able to point to my posts here as not sufficient for any VC interest. Maybe then, finally, the VCs will be clear on what their criteria are, in one word, traction. > 3) throw out outrageous $1T numbers, There is nothing "outrageous" there. Instead it is just simple arithmetic. If you are willing to entertain that there is a problem such that nearly everyone on the Web will find the first good solution a "must have" (e.g., like horse owners found the Model T), then consider the 1-2 billion users in the more developed countries, that is, where can get some good ad revenue. Assume that those people will go for the solution ballpark 3 times a week and each time see 50 ordinary ads (the user interface is interactive and iterative which is crucial for the core applied math but also good for showing users a lot of ads with, it turns out, quite good ad targeting). Assume the relatively good ad targeting. Then compare with the number of unique users, the daily user traffic, page views, ad impressions, revenue, and market cap of Google and Facebook. Done. Nothing at all outrageous. Or get a lot of eyeball time each week from nearly everyone in the more developed countries and, presto, bingo, can monetize to $1T market cap. Sure, a Web site for selling winter sweaters for puppy dogs on line might be an okay business, but, as a first step in picking a business idea or project, pick a problem that nearly everyone has, not just people who have a dog, cherish it, and want to walk it in cold weather. I tried to do that; dirt simple tactic. The problem? One I had and still have, daily. I believe that essentially everyone on the Internet has the same problem. So far there is no good solution. Why no solution? Can't get a good solution with only what is common on Sand Hill Road or in the Stanford, CMU, or MIT CS departments now. My applied math is nothing like what the AI/ML people are doing. I have nothing neural. I don't claim that my code is intelligent. No one screaming for a solution? Well, before the Model T, horse owners were not screaming for the Model T, either. Why not? They could not imagine something both better and at all realistic. Essentially every horse owner had the problem, and the Model T was the first good solution. Same for my work: Internet users have the problem but can't imagine something both better and at all realistic. Besides a big theme in information technology is moving ahead! E.g., in the last few months we have seen 14 TB SSDs, 16 thread consumer processors, and, from my ISP, download data rates of 100 Mbps up from about 10 Mbps over the interval, with the prices falling quickly. Main memory is going for ballpark $10/GB, one DIMM at a time, retail. Back there, 300 MB of disk storage weighed 750 pounds and cost $40,000. Big changes. Well, I have in mind another big change but will be not nearly the first. In lots of other ways, we have seen big changes. E.g., Amazon: They were an on-line book and record shop, but consider them now. How'd they do that? Well when Bezos started, no way was the Internet data rate high enough to permit him to send so many pictures of the products. Now the pictures are so good that shopping in brick and mortar sources does not have better product information, usually much less good. Big changes are possible. But, I can assure you, essentially no VC in the country sees my $1T and is interested enough for a phone call, an e-mail, or even a Tweet. E.g., USV just got a new investing partner who posted > If you’re a founder or potential founder building something transformational, I’m so eager to meet you. I’m on a new adventure and I’m ready to dive into yours, too. You can reach me at Rebecca@USV.com. So, I wrote her. Hear back? Nope! And I'm not holding my breath waiting! So, why do VCs post such garbage? Curious. Conjecture: They want to maintain the image that they invest in things that are "transformational" without mentioning traction. So, it's an initial tactic for getting "deal flow" and then deal negotiation. > 4) rile people up with arguments about VB.NET. What's to be riled up about? As a platform, I picked Windows instead of Linux -- lots of people and serious organizations have done that. On Windows, a key to nearly any software development of production software is the .NET Framework. Okay. Well, Visual Basic .NET can make apparently full use of the .NET Framework. Indeed, Microsoft MSDN Web pages documenting the .NET Framework, and I have about 5000 of them, are all or nearly all just the same for VB.NET as anything else, including C#. So, what's wrong with writing in Visual Basic .NET? I see nothing wrong. What is the much better alternative? I see no significantly better alternative. > Look, it's nice that you got some use out of Microsoft's frankly very irritating old developer tools. What? What "old developer tools"? I wrote in Visual Basic .NET which is an essentially fully up to date language on Windows and very different from the old Visual Basic. So, what are you concerned about?
ci5er 9y ago
Good luck. You may be a math genius, but you don't have a marketing or venture-formation bone in your body. (I do applaud you're choice not to take on VC money, though...)
- graycat 9y ago
  You have me suspected, accused, tried, convicted, and sentenced all in absentia and with no good evidence for doing NOTHING wrong.
  - ci5er 9y ago
    You are right. I was wrong. I am sorry.
calgaryeng 9y ago
Overconfident comes to mind....
- graycat 9y ago
  You have me suspected, accused, tried, convicted, and sentenced all in absentia and with no good evidence and while doing NOTHING wrong. No: Just a simple observation, e.g., much like horse owners needed something better, e.g., Ford's Model T which then became a "must have" for nearly every horse owner in the more developed countries. Essentially every user of the Internet has this problem everyday; currently there is no good solution on the Internet. The first good solution will be a "must have" for nearly every user on the Internet in the world. And my solution is not just good but excellent, better from being more powerful than possible with the techniques common on Sand Hill Road or in the computer science departments of Stanford, CMU, or MIT. My core technology is not computer science, AI, or ML, and is original with me. Current work in AI/ML can't get the solutions my work delivers. I'm able to do such original work; that's what my Ph.D. says and why I got it. The estimate of $1T market capitalization is back of the envelope, first-cut, ballpark but with some reasonably solid evidence: The main evidence is just the prospect of getting on average, say, 30 minutes of eyeball time three times a week from nearly every Internet user in the more developed countries (where can get good ad rates). Comparing with other successful ad supported Web sites, the $1T is reasonable enough. But for being speculative, it might be that, like cars where the Model T was not the end and not even the beginning of the end but just the start of the beginning, the usage and ad revenue might make the effort worth several times $1T. And there is a chance that the ad targeting can be come a big, new thing beyond just the ordinary views of ads for several times more. But that's speculative. Sure, cars are an example of how just improving on horses can become really big, e.g., create the suburbs. So are phones, PCs, dial up modems, AOL, CompuServe, and Prodigy, radio that became TV, computers, transistor based computers, microprocessors, and 14 nm line widths, aluminum, and more. Amazon is a big change, not just in retailing but even in logistics and the economy. So, with my project, it could become much bigger than the $1T. E.g., likely the current ad business would have difficulty supporting much over the $1T, but if go beyond ads, say, as Amazon went beyond just retailing and cars went beyond just replacing horses, and airplanes beyond what the Wright Brothers did, and satellites well beyond Sputnik, then could replace ads with something much better and get still more. But that's the speculative part. The $1T is reasonable enough, for a ballpark, first-cut, back of the envelope estimate. But, my applied math has some generality for taking in common data and putting out valuable information and could be applied to and be the core of something bigger than my current goal of the $1T. Then a solution, based on some new data and some rock solid applied math, complete with theorems and proofs (I hold a good Ph.D. in such work) based on some advanced pure/applied math prerequisites. The math is nicely typed into a document via Knuth's TeX. No, not LaTeX, just Knuth's original, long essentially totally bug free TeX. It's a nice applied math document. Am I going to publish that secret sauce? Heck no! A new, innovative, very different, but simple and easy to use, UI/UX that is interactive, iterative, somewhat game like, and for some users addictive and that helps generate the new data. The UI/UX needs just a Web browser up to date as of 10+ years ago. E.g., ASP.NET automatically wrote a little JavaScript for me, but I never wrote even a single line of it. That ASP.NET is the only JavaScript for the site, and even that is optional (I suspect it has to do with cursor positioning). The Web pages, then, are just dirt simple, send for <= 400,000 bits per page, with the JavaScript so load very quickly. There's no use of Ajax. The fonts are large; the colors easy to see (about 25% of men are partially red-green color blind but will be able to read my Web pages just fine); the backgrounds are all just white; the contrast is high. The pages are just dirt simple with just a few of the simple, standard HTML controls nearly everyone on the Internet understands; there are no icons, roll-overs, pull-downs, pop-ups, overlays, etc.; the screens just load, BAM, and don't jump around during loading. The pages will look fine on everything from a smartphone to a high end workstation. Production quality code in solid Windows Visual Basic .NET (no, NOT the old, simplistic, weak, unsupported Visual Basic 6.0), essentially the most powerful language for the .NET Framework on Windows, e.g., essentially just a different flavor of syntactic sugar compared with C#, fine, about as good as it gets these days for production quality software. The code is 24,000 programming language statements in 100,000 lines of code; so, right, is profusely documented in the code but also otherwise. Code running as intended with no known bugs and in alpha test. Code intended for first production through a quite significant company, say, $10 million a month in revenue with nearly nothing in opex. So, no prototype code. No minimum viable product. No stuck together with chewing gum and bailing wire. No refactoring needed. Instead, solid, production code. My first Web site, first use of SQL Server, first use of Visual Basic .NET, first use of .NET, ASP.NET, ADO.NET, my first UI design, my first Web page design, all production quality work from the beginning with no intermediate steps, prototypes, redesigns, refactorings, etc. I typed in the code myself using just my favorite text editor KEdit. I made no use of an integrated development environment (IDE). I had no difficult debugging. Timings show that the code is shockingly fast. The data base has only clustered keys, no foreign keys, and no joins. SQL Server is being used as not much more than just a B-tree data base system. In production, SQL Server is being lightly used -- the main data manipulations are elsewhere in my code. There is an off-line, batch program that reads the SQL Server data and does some initial processing later used on-line. It's a good project, a darned good project. There's just good planning, applied math, software development, etc. and nothing over confident. I didn't get into such work yesterday or even recently. I've been successful doing applied math and computing for US national security and business for decades. I know very well just what the heck I'm doing. It all worked just as I originally intended. There were some problems but not specific to my project. All the work unique to my project was fast, fun, and easy for me. Oh, by the way, I am 100% completely, totally, absolutely, permanently unemployable for any job above stocking shelves at Wal-Mart and likely even that. I'm a native born US citizen, never arrested, in good health, never filed an insurance claim of any kind, never sued or been sued, never convicted of any crime worst than a minor traffic violation, have not had a traffic violation in 10+ years, never used illegal drugs or used legal drugs illegally, etc. Yup, permanently unemployable. For a solo founder to build a company worth $1T is exceptional; same for $100 billion, $10 billion, $1 billion (the VCs hope for), $100 million. So, one might suspect that in some ways such a person and project would be a little different from what is quite common. So be it. A good thing about the Web is that the users don't know anything about the owner of the Web site; they might be a dog, wearing blue jeans.

a_imho 9y ago

I like the takeaways, but the no VC seems to be quite an exaggeration. They invested temporary resources (design, frontend) and a full-time salary (mine) into Snipcart, our developer-first e-commerce platform.

jacquesm 9y ago
That's not venture capital, that's a bit of boost at the front, and most of it not in money. They didn't say they self-financed all of it, just that they did not take on any venture capital.
- weego 9y ago
  OK so it's not VC in the narrowest of definitions, but there was capital investment by a third party so there are some language gymnastics involved to overstate the independent nature of the company.
  - jacquesm 9y ago
    There is nothing narrow about that definition, rather, you are trying to stretch what it means to take on venture capital to things it was never intended to cover. The language gymnastics are entirely on your part.
  - drharby 9y ago
    Moving goalposts?
- a_imho 9y ago
  Semantics, there is similar argument there without the exponential growth they are closer to a (micro) saas than a startup (discalimer, I've no idea how big Snipcart is) and their return is ~on par with the investment.
toyg 9y ago
Looks like Spektrum was basically acting as an incubator that spent some money on them. Your criticism might be valid depending on whether the incubator got some ownership rights (in which case it was acting as VC, albeit on a small scale) or not (in which case it was just an angel investor).

wordpressdev 9y ago

Being a solopreneur, I am a one man startup.

onion2k 9y ago

Sure, there are now dozens of e-commerce solutions similar to Snipcart. Yet there is 0 business like ours. It sounds like Snipcart are doing something very similar to Moltin (https://moltin.com https://moltin.com), which is a YC company.

tehwebguy 9y ago
My implementation of this was rejected in 2013. I should have put more work into my YC app + video!

ojr 9y ago

So it is 4 guys not 3 1/2 just because you don't work full time doesnt me youre 1/2 a person?

Sylos 9y ago
Right, but then why would you click on that headline?
- ojr 9y ago
  I clicked to figure out why a person was described as a half person but I didn't spend much time on the page and didn't figure it out, do you want to know what I ate for lunch too?
gggdvnkhmbgjvbn 9y ago
It was him plus the cast of two and a half men.

5_minutes 9y ago

Headline might be over the top but the essence here is a nice vibe.

zer0livemagic 9y ago

I'd agree that the fascination with VC money is not helpful. There are certainly verticals where you can only survive with outside investments (biotech comes to mind) or where scaling quickly is important, but no amount of money will save you if you don't have product-market fit.

anovikov 9y ago
Well, a good half of the ecosystem is tailored to eat away investor money and nothing else.