Slashdot Mirror

My story of when I was involved with an Apress book here.

Rich.

(My original comment up there was modded down to -1, but it's interesting to read the answers anyway)

You want to make sure that the layout is efficient and well localized, so that you're not stalling the CPU while it's fetching the next piece of data, but once you've used it you won't need it again for a long time

Exactly right! In the past I used C code to manually lay out OCaml objects to optimize access, so I could still use a nice high-level language, but get the speed advantage. The data size was approx 34 GB, stored in an mmap segment backed by a file. Section 6 in this document describes it.

Rich.

They went from using $1-2 thousand per week, to suddenly $2000 would get spent in 10 minutes between the hours of 1 and 2am. Google stone walled, denied, and finally did nothing for these small companies. I'm sure they aren't the only ones.

You mean: they didn't set up a correct daily budget, which is one of the simplest and most prominent adjustables in the AdWords interface. (In fact IIRC Google forces you to set a campaign daily budget whenever you set up a campaign through the web interface).

2) They are "forcing" adwords customers to have their ads listed on "link sites". that is a bad product, and if you are on adwords you are FORCED to have your ads listed there as there is no way to opt out

This is also nonsense. There are simple campaign targetting options so you can choose either where you do want your adverts to appear, or which sites you don't want them to appear on, or you can turn off content network entirely.

What you're really saying is that they had an incompetent campaign manager and a stupid business model (depending on a single form of marketing).

Rich.

You should be able to do both with OCaml.

http://sourceforge.net/projects/ocamldap/
http://merjis.com/developers/pgocaml/

If you had really complete freedom and were willing to try out something radically different from existing frameworks, I would suggest you would take a look at Ocsigen. It is based on the OCaml language, which alone implies a different mindset from traditional frameworks based on imperative languages. Some of Ocsigen's cool features:

• Extends the OCaml type-safety into the generation of XHTML. This means that producing valid XHTML is not only "nice", but actually enforced by the framework: your programme won't even compile otherwise!
• The entire site is seen as a programme where each public URL is a function. The OCaml type-safety is extended to forms and internal links, meaning there can't be any inconsistencies whatsoever.
• With database bindings such as PG'OCaml, you can extend the type-safety also to database access. Think about it: the compiler checks at compile-time if your programme is consistent with the database itself!
• Functional programming is very high-level, which means rapid development and happy programmers.
• It is fast. And by fast I mean really, really, fast. How would you like your web framework to generate native code whose speed is close to that obtained with C?

Sorry if this sounds like a sales pitch, but I would just like to point out that there are wonderful technologies out there, if people were just willing to take a step outside the trodden path.

I can't see that they've fixed anything very much in IE7. Our IE6 fixes used the "* html" hack, and that was fixed in IE7, but then we found we needed to include all the same fixes anyway. So now we end up having to serve two extra stylesheets.

I ranted about that on our company blog.

Rich.

Please, by all means write a forum BBCode parser that outputs valid XHTML that works under the application/xml+html mime-type.

The wiki I wrote does something analogous to that. (Not BBcode, but MediaWiki-ish markup which is similar). It's not really so hard for a competent programmer.

Rich.

if it depends on 6-8 MB for GTK

Our cross-platform "hello world" program including Gtk is about 300K compressed in total on Windows.

Rich.

The stupidity of the adverts is the responsibility of the advertiser, not Google... though Google does start to winnow the supremely stupid adverts - because they don't get many clicks. Adverts with few clicks will eventually get dropped by Google as being irrelevant to search users.

The system works by advertisers saying which words they want and any exclusion words. Only... most end users and a depressingly large number of agencies have no idea about using the exclusions. And Google makes it hard to identify the searches for which the advert was shown, but are irrelevant. Additionally Google will extend a search term if the advertiser has selected "broad match". Broad match is the default.. and it means that Google tries out similar search terms to see if anyone clicks. If users do click, then that search term is implicitly added. So the main problem in the "extensions" example is that the advertisers tell Google that they wanted the advert to appear whenever anyone used the word "extensions", but neglected to add "-file" or "-filename" or "-name".

Contrary to the impression given, Google does not offer behaviourally targetted adverts. Just search terms. Differences in adverts between two users usually comes down to geotargetting, and the complex effect of budgets.

Yes; I use Google and Yahoo to advertise for clients, and for our own stuff. IMO, it isn't an AI. It's a smarter system allowing more precision by advertisers, than Yahoo. Yahoo can't tell the difference between "car covers" and "car cover" (one is insurance and the other is a tarpaulin derivative) - Google can tell the difference and also see the difference between "covers car" and "cover car" (if it is meaningful) and may extend "car cover" under broad match to "auto cover", "car insurance" and "auto insurance".

I'm pretty sure that, with a team of programmers, I could replicate what Google AdWords does. Basically, user keys in search term, Google looks for adverts in the geoterritory that should appear against the search term, orders them by bid value (and click through rate), and finally checks that the advertisers budget permits another advert to be shown (by rate throttling if needed). If a user clicks, transfer to the advertiser supplied URL and rack up a fee for the advertiser.

No AI involved. Lots of fast web fronted databases, though.

Cheers, JeremyC.
--
Merjis: internet marketing by geeks : http://merjis.com/

But you're acting against a force of nature here. What's going to happen when people can transfer data wireless ad-hoc, encrypted over long distances? Or through their skin when they brush up against each other in a crowded train? How will the police protect you and your copyright monopoly then?

We are actually developing a training product. I can't say whether it will be like yours or not, because you don't give enough details, but it'll be distributed on CD and sold for £20 or so. This is because the market we are trying to capture (SMEs) still values stuff they can hold and they pay for. If people copy it and share it, you know what, I really don't care. It's a loss leader for us, designed to get us into the SME market which is absolutely the hardest place to be (so many of them, requires massive marketing budgets normally). It's designed to get us lots of small consulting contracts that we can mostly automate.

Sorry, but the world has changed.

Rich.

I run a small software company. The keyword here a is "small." Genuine small business with genuine employees making honest wages. After being tipped off by a customer, I looked at eMule and found that some of our software, which we sell for about 50% the price of our billion-dollar competitor, was being "shared" by 35 users.

I think it's unfortunate that you're sticking with a business model which requires artificial scarcity when for the last 10-15 years we've actually had a world of unlimited abundance.

Shame, but there you go.

Try releasing your code to everyone and wonderful things will happen (and you'll make a good deal of money through consultancy too, but that's only part of the fun).

Rich.

Admittedly not a great example of web design, but structurally it does the right thing. If you have firefox, view that web page and go to View -> Page Style -> No Style to see the structural markup (or just use View Source if you're comfortable with that). Of course we do use <div>, but only where it's essential. I would prefer to use it less, or even not at all.

Compare to this or this or this to see the overuse of <div>.

Rich.

It is also possible to call OCaml code from C. Don't believe me? Have a look at mod_caml - that's Apache (C) calling into OCaml CGI scripts.

My original call was for people to start using sensible languages which have support for bounds checking. You can write daemons and applications in OCaml perfectly well, and they will run as fast as C programs, using similar amounts of memory. The technology to exceed C in many areas has been around since the 1970s - for goodness sake, let's start using it!

Rich.

A blog which uses this attibute on link tags is far less interesting to comment spammers, so chances are the moderaters have to delete less spam.

Sadly this is not really true. I wish it were. In my experience - wiki spam - it appears to come from humans in China who simply cut and paste prepared text into the page.

On my wiki, the usual [http://...] syntax doesn't work - it doesn't generate a link (you need to use [[http://...]] to see a link), and my wiki even has an interactive preview which shows you that you're not making a link. Yet they submit the broken, non-working spam, and move on.

My sandbox pages have, for a very long time, been protected by a robots.txt file which would prevent them from getting pagerank anyway. There's even a huge warning at the top of the page telling them this.

We're not talking smart humans/bots here. We're talking people working in sweatshops.

Have a look at this history page to see multiple instances of Chinese spammers getting it wrong. I think there's only one case I can recall where the spammer actually bothered to get the syntax right, and it was still a waste of time for them.

Rich.

A blog which uses this attibute on link tags is far less interesting to comment spammers, so chances are the moderaters have to delete less spam.

Sadly this is not really true. I wish it were. In my experience - wiki spam - it appears to come from humans in China who simply cut and paste prepared text into the page.

On my wiki, the usual [http://...] syntax doesn't work - it doesn't generate a link (you need to use [[http://...]] to see a link), and my wiki even has an interactive preview which shows you that you're not making a link. Yet they submit the broken, non-working spam, and move on.

My sandbox pages have, for a very long time, been protected by a robots.txt file which would prevent them from getting pagerank anyway. There's even a huge warning at the top of the page telling them this.

We're not talking smart humans/bots here. We're talking people working in sweatshops.

Have a look at this history page to see multiple instances of Chinese spammers getting it wrong. I think there's only one case I can recall where the spammer actually bothered to get the syntax right, and it was still a waste of time for them.

Rich.

http://issues.apache.org/bugzilla/show_bug.cgi?id= 27550

http://merjis.com/developers/mod_caml/apache_2.0

Rich.

The MS and non-MS implementations are subtly different, but nothing major.

Have a look at the Javascript source. I do some clever stuff to measure the round trip time to the server and slow down updates if they are taking too long. Anyone know how to make updates run in another thread so they don't stop the browser if they're really slow?

Rich.

The MS and non-MS implementations are subtly different, but nothing major.

Have a look at the Javascript source. I do some clever stuff to measure the round trip time to the server and slow down updates if they are taking too long. Anyone know how to make updates run in another thread so they don't stop the browser if they're really slow?

Rich.

These languages face a Catch-22: until they're more popular, they won't attract enough developers to ameliorate the library situation, yet until they offer better libraries, [...]

Rich.

You could take a leaf out of O'Caml's book and make an interface to Perl.

I've also written an OCaml tutorial for people coming from 'conventional' languages like C, Perl and Java.

Rich.

but there are!
I learned the language online, (just as i learned Java from their very good tutorial)

they gave 2 good ones at the top.
the Richard Jones'
Ocaml Tutorial for people who know how to program 'normally'

There is also an update to Jason Hickey's book

I like OCaml because it Combines the power of functional programming, like (tail-)recursion, functions as an argument, with 'normal' programming language statements.
It doesn't force the "functional programming way" on you, like Lisp does, So, you can still use the If then, and While statements if you find them more usefull then a recursion.

And, it is quite fast!
both in development and in execution.
however, i have yet to find a way to well commenting my code.

If parrot can host ocaml then you will be able to call perl libraries from ocaml.

You already can: try http://merjis.com/developers/perl4caml/.

Rich.

... or even (my) free software cocanwiki.

Rich.

Another thing along those lines is what you get to use from other people. If you can use other peoples code, you don't have to worry about your own.
This is where perl is a big winner over everything else. Nothing else even comes close as far as number of available packages.

Except for languages like OCaml which can use Perl libraries as well as their own... and therefore have more available packages by definition. :p

I'd say ocaml is not suitable for general programming as evidented by its lack of general purpose libraries.

If that's your only issue, you can get round it rather easily by interfacing with another language famous for its extensive libraries but hideous syntax...

Maybe not Java, but someone should start an OS project with something like OCaml. You get most of the speed of something like C, with far fewer bugs. Oh, and there's no reason you can't compile your Java.

I would prefer a fast language without the usual pointer and buffer problems, like OCaml.

What a good idea. If only someone would write an Apache module... ooh, they have!

mod_caml

Rich.

http://www.merjis.com/developers/

Rich.

http://www.merjis.com/developers/

Rich.

Firstly I've written mod_caml which is, you'll be disappointed to know, an interface to the Apache API from Caml.

Secondly, OCaml isn't just a compiled language.

Thirdly, if you'd ever written a website you'd know that queries on the database dominate. Interpreted, compiled - it makes no difference.

A well-written CGI will be less efficient than a well-written servlet, especially under load.

This is plainly wrong.

Rich.

http://www.merjis.com/developers/mod_caml/docs.sht ml

Where possible I've gone for reducing the amount of code that you have to write in the common cases, based on a large amount of experience writing CGI scripts in Perl.

At the moment we're missing a fully integrated database layer, but that's coming soon (the code is already out there, I just need to pull it in and do the persistent database connection stuff).

Unfortunately I only get to work on this at weekends, but hope to have a Savannah page up soon so others can more easily contribute.

Rich.

[1] It's not really "scripting" as such. OCaml programs are bytecode compiled and dynamically linked into the bytecode interpreter which runs inside Apache. We also hope to get natively compiled loading working at some point.