Re: PHP on .NET DLR

On Sat, Aug 11, 2007 at 03:44:03PM -0400, Mike Schinkel wrote:
> K T Ligesh wrote:
> 
> > picked up in October. So maybe Andi knows something about why 
> > the PHalanger devs were moved to Ruby?
> 
> So?  You accuse MS of conspiracy against PHP. Clearly if they were against
> PHP they wouldn't have worked to get PHP running better on IIS.  My point is
> valid.

 It becomes more convoluted, and now definitely it is looking like a conspiracy. MS is promoting on PHP on windows but purposefully scuttling its development on .NET. Why? Again, give me some reason for that. I am now intrigued. Why is there such a antipodal reactions from MS regarding the same thing?

> > completely moved to Windows, since the Unix subsystem started 
> > working properly only in Vista. And the direction that web 
> > development is moving towards is _'desktop on web'_. So 
> > here's MS trying to deliver just that. I am definitely interested.
> 
> Personally, being a student of HTTP and the efforts of the W3C and
> technologies that have succeeded on the web and those that have failed, I
> believe efforts to move the desktop to the web are, in the best case, doomed
> to failure. In the worst case, I believe they are doomed to cause the web at
> large to fail, to cave in under its own weight.  Frankly, I think the former
> is more likely.

 To be frank, Web programming with css, table, div etc is a terrible thing. It is a completely structureless madhouse. As a developer, I prefer creating the entire gui by calling functions rather than by printing strings. The latter is what makes web development so difficult. Anyway, whether silverlight will succeed or not is irrelevant; it can't hurt anyone to have PHP on .NET nevertheless.

 OK, silverlight fails miserably. So what? PHP on .NET will still empower millions of PHP developers to explore new avenues, and come into contact with the new programming ideas that .NET library has. It can only be good.

> A more arrogant statement has never been made.
> 
> The problem is thinking that great devs are all that matters. Actually devs
> that can understand how to relate to the average user let alone average dev
> are those who have the ability to make great positive contributions. Great
> devs have no issue with complexity, whereas the average person is
> overwhelmed by it. Great devs simply don't know how to relate. Case in
> point, look at RDF vs. RSS. The former was designed by great devs, the
> latter by average devs.  Guess which one became significant?  The one that
> the average devs could understand.
> 
> Frankly I view one of my strengths being that I'm *not* a great dev. If I
> was, I simply couldn't relate to the needs of the rest of the world.


 RSS is not a language. The idea of programmer is to write extraordinarily complex programs whose very purpose is to hide the complexity from the end user. My day job involves building graphical interfaces--even though I personally do not use them--to hide the complexity of the system from the end user.


> 
> > Software is not 
> > writing 100 lines of code and feeling clever about it. 
> > Software is about building and maintaining extraordinarily 
> > large projects, 
> 
> Some software is about large projects, but a lot more software is about
> small projects and more importantly, components. All open-source software to
> speak of is either comprised of many components or is such a component.

 Just previously you complained how difficult it is to stitch together drupal, wordpress etc. Why? Because there is no single person who can handle that kind of complexity. Complexity is the key. And complexity exists so that it can hide itself from the end user. That's programming. You make your programming incredibly complex so that your user doesn't have experience it.

 Drupal, wordpress etc has decided to keep it simple. What's the end result? Complexity has been passed on to the end user. This is one of the absurdities of Open Source. Programmers are saying: We want simplicity, so we will throw all the complexity at the end user.


> 
> > and what you need is not simplicity as such, 
> > but rather the ability to abstract.
> 
> Hah!  Tell that to the people are are *not* using RDF.   RDF is all about
> abstraction. Too much so.  Simplicity is winning the world over today as a
> way to manage information overload, in case you haven't noticed. Google,
> iPod, and more.

 RDF is not a language.

> 
> Languages are evolving. C-based languages are evolving too, but C-based
> languages may be too stuck in the past to effectively evolve. They has too
> many concepts that are incompatible with newer, more elegant, and more
> simple languages.
> 
> BTW, one incredibly beautiful language I recently learned is from MS:
> PowerShell. Its simple but scales its complexity well. OTOH, its language
> design needed a clean sheet; it wouldn't have worked on the back of a legacy
> C-based language.

 Powershell is not meant for writing applications. But funnily enough, I love powershell, because it is pure C#. You gotta love MS, how they are screwing over their VB programmers. No more VBscript. Now even Windows shell scripting uses C. Ha.  That's another reason I have started liking MS. 



> > And python a simple language? Are you joking? 
> > I don't think even the die hard python fanatic would claim 
> > that. 
> 
> I guess you've not read any of the books on Python programming that I've
> been reading lately. Alex Martelli discusses how Python focuses on being
> simple to understand. BTW, simple language does not mean "simple under the
> covers", it means "simple to use."

 To be frank, python focuses how to make it easy for the compiler writer to get away with least amount of features. Python doesn't even allow you to embed variables in strings. How insane can it get? And you need to add 'self' to every method in a class. Python existed because perl was really stupid. I am happy that because of ruby, python mania has died down.


> 
> And yes, Python can scale in complexity, but it is very easy to get into and
> it is very easy to do basic things in. It is also very easy to read, much
> more so than PHP (I'm thinking of Drupal modules at the moment.)  And Python
> has the ability to define syntaxes for domain specific languages (as does
> Ruby.)

 Sorry. The general perception is that PHP is a bad language but simpler than Python, and that's why people perfer it. If you are so much against braces and C heritage, you might like it, but python is generally not considered as simple.


> 
> For example, I just watched a presentation of SQLAlchemy the other night.
> Brilliantly simple; I could teach a 12 year old how to use it.
> 
> > Python is supposed to be an elite language. 
> 
> By whose definition?

 People who use python. 


> > In fact, as 
> > everyone points out, without Rails, I sure none of the code 
> > monkeys would ever touch Ruby, and run away scared from its OO.
> 
> I highly doubt OO is that off-putting these days. It was in 1990 when i was
> first teaching it, but it is not today.

 That's the idea I get reading comments posted everywhere.

> 
> >  MS is gaining hard on the server arena. Being a server 
> > services vendor, we have no choice but to fully support 
> > windows or lose a huge market. IIS powers 31% of the websites now.
> 
> But IIS supports PHP as I quoted already. What's the problem?

 Why stop there. Why not take the next step. As I said, for backend I will be using PHP/IIS, but what will I do for the silverlight frontend?


> 
> >  To me they look good enough, especially considering their size.
> 
> But you are not on the inside so your view is not as clear. '-)

 It is irrelevant to me. I don't want to take any chances.

> 
> >  Yes. Obviously. But they learned, and there is still life ahead.
> 
> We'll see.

 As I said, this is not a discussion of whether silvelight will succeed or not, it is about giving your PHP developers, the ability to develop apps in .NET platform too. If you merely consider that topic alone, then there is no ambiguity there: IT is a really great thing.

 So now the next step, what can we do to make this happen?

 Thanks for the nice conversation.