Webolio: My own views on the web

On Partnerships…

Simo — Wed, 10 Sep 2008 21:26:30 +0000

Building partnerships is an exercise web entrepreneurs should master. When venturing with someone into a new business, be it an investor, a co-founder or an employee, make sure you are with the right person:

Trust is important. If trust cannot be established for any reason whatsoever, ditch this relationship and move on.

If you get frustrated too much too often because your partner cannot appreciate or understand something you’ve done or a decision you made, then it is not worth investing any more time in this partnership…

…which leads me to this: always have an exit clause on your partnership agreement. Not having one leaves the door wide open for legal conflicts in case stuff doesn’t work out.

Always consult a lawyer before starting a partnership. NEVER do this alone. This will save you countless dollars and trouble down the road.

A partnership is like a marriage and remember, only fools rush in.

My Cari.net dedicated server is down..and so is cari.net!

Simo — Thu, 13 Mar 2008 17:16:55 +0000

Don’t want to be bitter but I am starting to get fed up with cari.net…The dedicated server I bought with Cari.net 2 weeks ago has been down on three occasions so far!!! that’s a lot… lucky I only use it as a dev environement at this point. Imagine if that was my production server.

What sucks the most is that I wasn’t contacted at any point regarding a scheduled/unforseen outage on all three occasions. The last one was just the worst of all, my servers has not been accessible for the pat 15 minutes?! so is cari.net anyways!

That without counting the fact that it takes me sometimes 10 seconds to open an ssh session (I have a 18Mbps connection at my home office)

Let’s hope these issues are addressed promptly and a valid explanation is provided! otherwise, my cari.net adventure will have to end right here.

PHPMailer and smtp on 1and1 shared hosting

Simo — Sun, 02 Mar 2008 18:08:15 +0000

I use phpMailer on my 1and1 shared hosting account. Recently, I could not send email anymore.

I tried debugging and this is the error that phpMailer throws:

Language string failed to load: connect_host

After googling for a solution and trying different smtp servers, accounts and smtp ports, I decided to switch to sendmail and it worked like a charm! All I needed to do was replace $mail->isSmtp() with:

$mail->isSendMail()

sendMail is located in its default location on 1and1 servers : /usr/sbin/sendmail, so no change of settings was required.

Conclusion: 1and1 probably closed its outgoing smtp ports on its shared hosting servers. Consequently, if you use PhpMailer, don’t use smtp mode anymore.

Getting Hover to work on IE7 with other tags than anchor

Simo — Sat, 12 Jan 2008 18:53:25 +0000

Before IE7, hover used to exclusively work with the anchor tag. On IE7, unless you make your HTML page follow a strict html4/xHTml1.0/1.1 dtd it will still not work… So you actually need to add this to the top of your html page:

a more detailed post regarding this is in:

http://www.bernzilla.com/item.php?id=762

Voila!

reading docx files using openoffice

Simo — Fri, 04 Jan 2008 19:59:13 +0000

Found this today… I tried using OpenOffice to open a docx file (which is a Microsoft xml format of their doc files) and this was of a great help. It basically tells you how to enable OpenOffice to support docx files. Pretty neat:

http://www.blogmanno.com/?q=node/64

Building an Online Service: Golden Guidelines to Remember!

Simo — Sun, 08 Apr 2007 03:19:18 +0000

Tens if not hundreds of web services appear on the net every given day. Some of them do get the attention they deserve, yet a lot of them don’t. Now every service is different and causes of failure vary accordingly of course. Having tested many of today’s web2.0 services, I can confidently say that most of these services don’t fail because of a lack of functionality, but mainly because they fail to respect a set of basic guidelines that are often overlooked:

. A solid “raison d’etre”: the most obvious, yet the most important. Is there a market for your service? Does it solve at least one problem right and better than other services? Did you talk about your web service to friends and acquaintances with the same problem? Did they show vivid interest in your idea? Your online service really has to have a strong basic reason to exist.

. Easy (or no) registration. I think this is a fairly well understood practice nowadays, but there are still some services out there that ask you for your birth date and phone #. Yeah, yeah, sure you have to know your audience, but that’s something that can happen when you have reached the 100,000 user mark, and when your service has become so viral that people are willing to sacrifice a few minutes to have their account/page. Better yet, you can always organize surveys in exchange for some incentive when you need that data. And believe me that will dramatically reduce the bias of your data, since those users will be much less prone to entering fake information. If anything, users should be able to use your service without registering.

. Easy use cases: You have, on average, about 20 to 30 seconds to capture user’s attention. Most users will get bored (and probably never come back again) if they aren’t impressed in that short amount of time. Moreover, if they’re using another similar service, they would at least want the same level of simplicity and a more user-friendly environment. Not functionality, but user friendliness. Extra functionality is good but that’s currency you will use to build user loyalty, which by definition comes with time. Bottom line, you have a very short amount of time to make a good impression and it has to be Good! You don’t have a second chance. Keep things simple and you’ll be amazed by the results.

. “Light” home pages / landing pages: don’t overload the initial page that a potential user will first use on your site; although bandwidth nowadays is not a problem for most internet users and hosting is cheaper than ever, still it is always a plus if your page loads quickly.

. Professional web interface: brand your pages well. Everything on the web has to do with design and layout, therefore, a visually pleasing interface is a must. That should not go against simplicity and loading speed though. Hire the right designer(s) to get this piece done if you don’t have the right skills. This is an area where you cannot afford to be cheap.

. Nicely trap all errors: In case your service is still beta and a bug appears in your application, graciously let the user know that an unexpected error occurred and that the development team will promptly look into it. Surprisingly enough, experience taught me that even the most average users have an idea of what a bug is and are forgiving in this respect. They will try again later provided they are treated with respect. And by respect, I mean they are not shown some cryptic message that even developers would have a hard time deciphering. Also make sure all these errors are properly logged and sent to the development team ASAP.

. Release gradually: Gradually unveiling your product to the outside world guarantees that you solve problems that may arise on a smaller scale, and also refine your product as you go with users’ feedback in mind. First you start with a private beta, where you use a handful of users to test your product (these are people you generally know fairly well- some of them might be members of your the team) . You use their feedback to get your product a bit more robust, then you release a public beta and test your product on a larger scale. Once you reach the right level of confidence about your product and add some of the requested features that your design team has overlooked, you can release it the public. This approach will almost never fail unless your product has really no reason to exist.

Startups: How much equity should be given when “nothing” is at stake?

Simo — Mon, 26 Mar 2007 05:45:49 +0000

Equity as the sole form of compensation is a recurring issue in today’s web2.0 boom. Many software startups need to hire to grow, however many if not most of them don’t have enough funds to hire the necessary talent for the obvious reason that: everyone wants cash.

Now let’s say you found someone willing to sacrifice a few months without salary. How much equity are you going to give? this brings up the issue of valuation. Equity should be correlated with the value of the company and of course the nature of service the new hire will provide.

In my case, since I have nothing but scattered ideas about how my product will be, I would say that at this stage, the valuation of my company might not exceed 3 to 5 hundred grands.

Now, my partner-to-be who happens to have much more experience than I do in startup related matters, wants to gradually get involved in my company. He is not ready yet to commit fully to the startup but he is willing to participate in some initial aspects of product development as well as serve as a mentor/advisor to me.

After several days of thinking, I decided that 4-5% should be good enough.. Normally, most startups give their advisors a little more than .5% – 1%. However, at the stage I’m at right now, I need his expertise more than that 4-5% equity. So I hope he’s going to accept. Of course, the percentage my advisor will get will be vested over a couple of years, and as soon as money comes in, I will start paying him for his time which should normally be a couple of hours a week.

Sure, if I had a prototype in place and had been through one financing round, things would have been different, but given the current state of things, I will be more than happy that someone of his level of expertise and interest in my product is on board.

In my view, there is nothing wrong with raising the bar a little higher than the advised 1% advisors usually get especially when your company is at its infancy. Several posts on the web refer to sub 1% percentages (Check this one in particular), yet they do say nothing about the valuation of the company and the value the advisor will bring.

for that particular reason, I believe there are several things to factor in when deciding what amount of equity a partner (or advisor in this instance) should get when joining a new venture:

Can he brainstorm with you on product related issues?
Can he participate in the design?
Can he be an early beta tester?
Can he provide you with legal advice?
Will he be willing to introduce you to his connections?
Does he have any previous startup experience?
Does he have any business background?
and last but not least, is he available to listen to you when you need him?

Will keep you posted regarding what my partner-to-be will say about my offer, and more about my startup.

Conversion d’une valeur numerique en texte francais (code for converting a numerical value to french text)

Simo — Mon, 26 Mar 2007 01:45:47 +0000

Pour ceux qui en auraient eventuellement besoin, j’ai uploade un snippet pour la conversion d’une valeur numerique/monetaire en texte. C’est un snippet que j’ai developpe il y a une dixaine d’annees (le temps passe vite!) et que j’ai decouvert en faisant le tour de mes vieux fichiers.

Je ne sais pas si ce sera d’aucune utilite pour qui que ce soit maintenant avec toutes les librairies qui existent, mais bon le voici tout de meme. Notez qu’il est ecrit en Basic (et le niveau de sa programmation est basique aussi ) donc il peut etre mis dans une macro VBA.

Installing Windows XP using a non bootable CD and no floppy disks

Simo — Tue, 20 Mar 2007 21:41:11 +0000

Yesterday, a friend of mine asked me to install windows XP for him and take care of finding and downloading all the necessary drivers to his machine.

He provided me with a copy of Winxp Pro Cd. However, the CD was not bootable and there were no boot floppy disks with it.. So I found a pretty simple way to do this. It took me a while to find the solution, so I feel compelled to share it with those who might run into the same problem: (you need another working computer with a CD burner to do this by the way)

1: Get bootcd . Boot CD lets you create a bootable CD that has the necessary utlities to get your computer up and running for a windows install. It also allows you (among other cool things) to make your C: drive bootable so you can later on launch the windows install from the CD.

2. Uncompress the .zip file. The file contains an ISO file that you can use to create the Boot CD using your preferred CD burning tool. (I used Roxio)

3. Boot up with the CD you created (don’t forget to activate booting from the CDROM on your BIOS settings)

4. You will be given several options as to what CD driver you want to use.. press enter (first option offers support to most CDROM drivers)

5. Switch to the R: drive (all the utilities provided on the CD can be accessed from there)

6. Now depending on the state of your OS installation, you might wanna play around with fdisk or some other utilities to create a FAT32 partition, or format an existing one (refer to the README file that comes with Boot CD for the list of utilities and what they do). let’s assume for simplicity that you need a fresh install, call fdisk to create a new partition (and delete the existing one(s) if any), then format it using “format”.

7: Execute a utility called setuphd so all the basic system files are transferred to the C: drive. This will make C: bootable and make the CDROM recognizable from the start.

8. We’re almost there. Reboot your computer (remove the CD now or change your BIOS settings to boot up from the Hard disk)

9. Insert the Windows XP installation CD and type R: to switch the CDROM.

10. Go to directory R:\i386, then type “winnt” to launch the install.

11. At this point windows installation will take over.
Quick note: You might wanna convert the FAT32 partition (if you formatted the whole disk as FAT32) to an NTFS one. The Windows installation program will prompt you regarding this. Or you can keep it until all is installed and use the NTFS conversion utility that comes with your XP.

Ok, so here is my quick and sloppy guide to installing a non bootable Win XP CD on a computer with a defective or no OS. Hope it’ll help someone out there.

O’Reilly’s Safari

Simo — Mon, 19 Mar 2007 02:21:59 +0000

I did come across it a couple of times before, but I never thought I was ever gonna need it this bad until last night; I was quietly working on this project of mine which has a CORBA component. Being an absolute beginner in the field (though I did some RMI in the past ..far past) I kept on jumping from one page to the other looking for a solution to my headaches..memory management in corba sure is not a piece of cake. Anyways, short of solution and after a couple of posts to some forums that yielded no answers, I could not wait anymore and decided to get this book on CORBA and C++ which is considered by many as The book for serious corba programmers. Then I stumbled upon Safari…for $39.99 a month, you get access to the entire O’Reilly collection of books for geeks

I eventually managed to solve my problem quicker than I initially expected. Book navigation is a breeze, plus you get relevant suggestions to other chapters in related books as you page through the book.

I might have waited a couple more days for someone to get back to me on those forums, or just continued playing around with keywords on google hoping that I find a post by someone who has had a similar problem AND who has solved it AND published the solution..but that would have been a drag.

Though not a big fan of reading stuff online, I think this kind of service complemented with a good search service like that of google book search has a bright future.

Bottom line, not everything is on line… I definitely think making books available on line is something extremely useful, now the biz model around it is yet to be tuned. 40 bucks /month is still a lot of money, I think.