Older blog entries for gnutizen (starting at number 26)

I am installing the software for my new HP Photosmart C5180 on my Windows XP. What the hell is it doing? It's been installing junk for 15 minutes and is still going. All I want to do is print out a simple page of ASCII text.

"Installing Microsoft .NET Framework" Overall Progress - Installing (Step 3/4) 40%

Blah.

Slashdot mentioned that Google had release a free (depending on it's dependence on the included aspirin package, and the status of that) OCR program. I tried it out on an image that clara and gocr had done, it bailed on me. I just ran one of the gocr test images with it and it did OK (although it had real problems with numbers, and seemed to split one letter into multiple letters, e.g. W -> VV). It is written to C++. I'm glad they released it, although I don't think it is the complete free OCR magic bullet yet, more work needs to be done on it, or it needs to borrow (or be borrowed) concepts and code from other OCR programs etc. I want to contribute to free OCR but I don't have the time so much.

Right now I am dealing with using things such as web browsers via Windows Remote Desktop and UNIX's X-Windows and it occurs to me that Remote Desktop is much faster than X-Windows for this sort of remote stuff. I wonder if any speed comparisons for this sort of thing have been done.

I have been looking over the web sites which show what kind of hardware is compatible with Debian (or Linux in general, or free software even more generally). I'm looking not only at the standard motherboard and hard drives, but things such as scanners and so forth as well. I have never done any kernel hacking before, but I might do so on the machine I will buy. I am not looking to go cheap-o either, I want a kick-ass Debian desktop which I will use for my entree into kernel hacking as well as just being my general desktop. I have caught the Debian bug and will not install non-free Java and the like on it - if it's not free I won't use it. I've seen some of the larger Debian/Linux hardware compatibility sites and FAQs, if anyone has any comments about all of this I'd be happy to hear about it.

Wage slavery

After a blissful year of working for myself and being poor, I am becoming a wage slave again. When I was in my 20's, my thoughts were on making a lot of money and getting stock options that would make me rich. Nowadays my thoughts tend to be on working for myself. I know if I worked for myself and made $70k a year, I would be very happy, I'm not sure how much less than that I could make and still live without having to worry about money.

Anyhow, me and my partners business is off and running - we are selling on e-Bay, our osCommerce site is up with its modules, we are in Froogle, we have all the corporation papers and have accounts with suppliers. Hopefully, with all the time I will be spending on my new job, all of the groundwork is done, and I can run the business on the weekend, along with my partner and others peripherally involved. An influx of capital will be helpful in some respects.

As I said before, my perspective has changed since my early 20's. Time is more valuable to me than money nowadays, although it is not easy to find a 40 hour a week IT job, everyone is hellbent in expropriating more and more surplus labor time from me and everyone else. I would gladly take a large pay cut to work a 40 hour job, but that is quite difficult to find nowadays. I am quite unhappy selling myself back into wage slavery, but hopefully the capital influx etc. can get me going on working for myself. It's unfortunate the workers of the United States have to spend so much of their lives working to achieve their own freedom from the bonds of wage slavery, with few even being able to even achieve this. Of course, in terms of individual struggle this will just get worse, only a long campaign of organized struggle will put an end to this.

Mucking with Linux modules

It took me so long to get my CD-ROM burner, DVD player, and USB wireless ethernet adapter (which uses an external kernel module) to work for Linux 2.4, that I haven't gone to 2.6 yet. I'm too busy and don't want to go through all of that again. But anyhow, I want to receive faxes on my box and started mucking with that stuff. I compiled the related kernel modules. Now I'm reading that my modem has no free as in speech, or even free as in beer Linux drivers. I hope I don't muck with this for hours more only to find out there's no Linux support and I can't receive faxes. In other news, my roommate's Windows Internet connection keeps conking out for no reason, so I guess it could be worse - I could be using Windows.

So anyhow, my journey into the world of e-commerce continues.

This weekend I signed up with Dreamhost. I was going back and forth of whether to do the expense for a while, but it is only $120 for the next year and I get a lot of stuff (web hosting, mail, ssh etc.) so I just jumped in and did it. Then LA has a blackout the next day! But they recovered nice and did a quick fix for an NTP problem I had complained about, so I'm happy.

I set up osCommerce there. They have a pushbutton way of doing it, and it was pretty easy. Hopefully I will be able to delay the need for SSL and a unique IP. The Post Code thing bothered me so I changed it to Zip Code and made some other changes. So now I have to figure out what I'll do for the credit card backend (we'll probably use Paypal to begin with), and perhaps I'll sign up for an account with USPS. With that done we'll do a test purchase or two, then start looking for real customers. No big rush though - I haven't even set up any mail yet, all of my e-mail is still going over Yahoo Mail for this.

Most of my sales have been over eBay thus far. Later on we'll try to drive sales to the web site through various methods. I've been analyzing sales on eBay for my little niche. I didn't like their watchlist system so I designed my own. My system uses a MySQL backend. I grab information on auctions with a PERL script. I display the results via PHP on my private Apache web server. I'm just trying to get an idea of what sells, and for how much. If something sells consistently, at a price I can make a profit on - I buy.

Which brings me to suppliers. Finding and dealing with them is a bit of a pain. We are low budget, because that seems like the smartest thing to be, we want to put off spending as much as we can - these people sometimes want to do a credit check on you before they even tell you what products they sell. They expect me to have a fax number as well as a regular number. Also, eBay is cut-throat - in my little niche, I can't offer a competitive price on most of the stuff that sells on eBay. We have found a handful of items we can make money on though. As we find more suppliers, buy in more quantity and so forth we'll probably do better.

I noticed eBay has an API interface which allows several thousand free queries of their database per whatever. I signed up for it but have been too busy with other things to look into it much.

Doing sysadmin work for the dot-bombs and later Fortune 100 financial companies, I made over $80k a year. I have been working full-time on this for the past few months, and am looking forward to getting up to the point where I'm making $20k a year. Yes, it is less money, but I do not have to answer a pager, I do not have to worry about being laid off, I do not have to listen to a boss and all of that junk. I do expect to eventually get back up to my former salary though, and do it working for myself, which, now that I'm in my early thirties, is more important to me than working for someone else and making more money. I think over the long run, you only make more money when you're working for yourself anyway.

I am doing a lot of PERL programming now. I've written short scripts in PERL for years, and wrote one very long script a while back, but am not much of an expert. I'll have to learn how to deal with the things needed once scripts get long enough so they don't become spaghetti code. My code was looking like spaghetti code until I broke it into sections - get information, parse information, print information. Next step is putting the information into MySQL, joy. Anyhow, this is taking less time than I thought it would, which is good. Once I finish my information grabber PERL script(s), I will write some PHP (and possibly PERL) scripts that pull this information from the database and put it on a web page. Luckily, CPAN has a lot of the functions I need written already, like HTML parsing. Splits, shifts, pops, and the ability to do sed/regexp stuff easily like this:

$variable =~ s/"//g;

is quite handy. I would probably write my scripts to view this information in PERL as well, but I feel PHP is better, and simpler, for this type of interaction with MySQL. Anyhow, this is all elementary stuff I suppose, but I'm just surprised how quickly and easily I completed my task. If I was a real PERL wizard, I'd probably have done it even faster.

I'm looking for a place to host a web site, and want to know if anyone has any recommendations from their experience. The ideal web site would let me have ssh access, would have e-mail, ftp, Apache, MySQL, PHP/PERL, SSL-capable, OScommerce capable etc. It would also have good uptime, decent speed, and that sort of thing. A lot of stuff I've been looking at is in the $10-20 a month range which is what I'm looking for. The ability to host several domains for this price would be a plus. Any suggestions? The one I've been looking at most is Dreamhost. Another I've been looking at is Liquidweb (possibly via what I believe is a reseller of it, addaction, or maybe directly). Does anyone have any thoughts of these two, or have been happy with an alternative site? Especially if it runs Apache, PHP and MySQL and has shell (ssh) access.

In other news - Sourceforge finally fixed their damned statistics. I did some nice recoding of Gnutizen in February, but have been busy since then. Gnutizen is still alpha since it can't share. I wonder if I should allow it to be an ultrapeer before going into beta - I guess the Gnutella community would demand it. There's not much of a reason to use mine other than others anyhow, although I wonder if any Gnutella clients work on a UNIX command line anymore, I haven't looked for a while. I know gnut doesn't. It's a good thing I cleaned up the code before trying to implement sharing anyhow. Geez, it was a mess. Code Complete was a helpful read.

autoconf and automake - ugh. They remind me of what sendmail used to be like.

I rewrote a lot of Gnutizen, especially the Gnutella packet parsing section. Relative to the old sections my functions are much more clear and logical, I look back at the old ones and say, what was I thinking? Anyhow, I have been working for a while to clean things up before I add new functionality, and I am about finished with all of that, which is good because I want to add new functionality like the ability to share files.

One of the main reasons I'm writing Gnutizen is I'd like to see some of my ideas for peer-to-peer advancement realized and one way to do that is to write a peer-to-peer program. Not so easy - especially when the protocol keeps changing! With the advent of file hashes for Edonkey, Gnutella and Bittorrent however, I'm beginning to see that I might not even have to write all of, this, in an "art of unix programming" way, I might just be able to write what I want and then "pipe" it to another program, in the same manner web browsers launch video files with a fork to a video player. It would be real simple for me to get some of this done, so I won't even go into detail here. I'd like to have a look at the basics before thinking about the harder architectural problems.

29 Jan 2005 (updated 29 Jan 2005 at 11:54 UTC) »
Bolsh: I do not even agree that the assumptions behind the "law of supply and demand" is accurate in your reference to Robert. Price is not a "mechanism of information transmission" by any means except in a negative sense - if some corporation produces some commodity that people do not want to buy (e.g. exchange for something of equal value), then it may realize it has wasted its time producing something no one wants - something that happens all of the time in economies such as the US's. But in that case, information is only transmitted in cases where you're losing money making and selling something. Thinking that numbers such as prices, or even the objects themselves have any inherent information or value in and of themselves is fetishization. All value is simply congealed in a commodity by a homogenous measure of labor time. Of course, in many ways, rent does not even fall as easily into our modern economy's rules since it is a holdover from the days of feudalism - we even call the people we pay rent to "land lords" as they did 1000 years ago. I live in New York City and rent control has worked fine here and I have benefited from it. There is no sense of a lack of supply at all unless you absolutely must be one of the million and a half people packed into the 30 square mile island of Manhattan. And people I know in San Francisco had no trouble finding apartments which were decently priced in the Mission or Oakland even at the height of the boom. I think the real question is which is crazier - rent control, or an economic system with the kind of insane boom and bust that sends people into San Francisco like madmen before sending them fleeing? Rent control is a minor blip on an insane economic system which veers between massive overproduction and recession, to mention just one of its negative qualities. One would think what is wanted is that someone working 40 hours a week at a useful job can afford to keep a roof over his head, and the economy should be geared to that, instead what is being suggested is that we have an economy that is not there to serve us, but that we exist to serve the economy and it's crazy quirks as it veers about. Of course different people look at this differently - I rent, but I'm sure a landlord would be of an opposite bent, there is not one correct view of how this works, but two conflicting ones. If there wasn't, the renters would have probably have "seen the light" of the landlord's soi disant wisdom decades (centuries) ago.

I just skimmed through Code Complete, 2nd edition and liked it. I just found the big bug I have been hunting down recently - the format of what was coming through over my socket changed. The "program defensively" section is obviously most needed when one is listening to sockets over which almost anything can come. The section on functions was what I needed to read the most. I also browsed through the K&R book again reading about functions. I've also been playing with valgrind, ddd, gprof, splint and other tools more. My code is getting better and less buggy, hopefully it will become more so as time goes on.

23 Jan 2005 (updated 23 Jan 2005 at 04:51 UTC) »
My Wintendo

I wanted to check out Exeem on my Debian, so I downloaded WINE and loaded it up. At one point, I had a menu scroll down and moved my mouse and the menu just hung there, as it does in Windows, and even followed me around across workspaces. *Shudder*. It brought back bad memories of the days I had Wintendo on my desktop. Of course, this is just one of the things that goes wrong all the time in Windows, and WINE faithfully emulates it. And people complain about GNOME. I wonder how Debian and WINE handle Exeem's Cydoor's spyware. Pretty well it seems. Spyware and adware have gotten worse as time has gone by, I had one invade a Windows box (from Overnet) that took me a week to remove instead of the usual minutes. It was ftp'ing stuff and everything.

17 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!