Seva Lapsha at Work & around

Making Collective Decisions

Seva — Tue, 13 Dec 2011 08:49:28 +0000

While preparing to shrink up and replace my single role of Lead Software Architect with a Software Architecture team -

Decision Making

Making architectural decisions might seem easy and fun, but can have multiple consequences which can affect state and integrity of the software, the architecture team, the R&D teams and the whole company. We aim to make the decisions in the most responsible manner with taking all the relevant factors in consideration. Sometimes we do consult wide community of developers and PMs in early stages, sometimes we do that on later stages, but what’s very important, we never attempt to dictate our decision without majority consensus.

The common flow of taking a decision is the following (CRAFT):

Collect information about the subject from provided private and public sources. If necessary, use in-company resources reuse their existing experience.
Research the gathered information and prepare a draft of architectural opinion with a list of alternative solutions and their advantages and disadvantages.
Acknowledge the opinion of all the interested parties. Accept and react on all the provided input.
Finalize the decision and make it effective. Make sure it’s acknowledged, accepted and adopted by all the relevant parties.
Track the execution of the decision. Learn from the experience, reopen the flow if necessary.

Simple No-Framework Object Oriented Multi-Layer MVC Application Example

Seva — Tue, 30 Aug 2011 11:59:26 +0000

Recently I’ve created the subj. It’s a memory game, written in pure PHP/HTML, called “PHP Memory”.

What was important for me to demonstrate is the next principles:

Even if you don’t use a framework, you should write modular code with appropriate architecture.
Despite that, the application design and code should match its required functionality with no overhead of unneeded patterns and abstraction or preparations for future enhancements, unless they’re planned.
Nevertheless, the code should be readable and maintainable.
The bottom line is – developing a framework versus developing an application are completely different tasks. The strategy and tactics should vary very much. Of course, developers should learn from the frameworks source code, since they accumulate collective experience of great coders, but the ideas implemented in frameworks sometimes not required or even harmful while developing an application.

The demo and source code can be found here.

cURL HTTP1.1 empty POST bug

Seva — Tue, 02 Aug 2011 22:47:35 +0000

Today we spent almost 2 hours on a weird discrepancy between our development and staging environments. It’s pretty rare, that I experience such low level issue, thus in my opinion it’s worth mentioning here.

Last weeks we were busy developing integration to a new data vendor. Everything went well until we deployed the application to stage.
Suddenly we started receiving HTTP status 411 on one of the calls. Since we work with cURL library, which we believed is stable enough, we thought the problem is somewhere between the source code and environment configuration.

Later we found that the same request gets accepted if sent from a client other than cURL (e.g. chrome-poster). The unique about this request is that it’s sent with POST method (the vendor’s strict requirement) but the content body is empty.

In the end we discovered that newer version (since 7.20) of cURL interprets missing body as a negotiation request – sends Expect: 100-continue header and Content-Length: -1.

So, the immediate solution was to send empty content body (zero-length string) to cURL, which aligned the behavior in all the environments.

On the way we discovered a useful option CURLINFO_HEADER_OUT, which enables a possibility to further retrieve the headers sent by cURL to the remote host. From now we use it in our error handling mechanism to trace the sent headers as well.

What can we conclude from this story?

Try to synchronize the software of all of your environments. If possible, use exactly the same version OS version, libraries, tools, etc. It’s very easy if you host your applications on VPS and use VMs for development and staging servers. The least convenient case is when you have OSs of different architecture and (e.g. Windows for development and Linux for production).
Don’t underestimate importance of error handling. Find the optimal level of handling for your application, which will be easily extended and configured.

Hope my sincere advices will help you, my dear friend, to save your precious time.

Occasion

Seva — Thu, 02 Jun 2011 23:15:11 +0000

Today I’ve been told the second time in my life that I look Abdullah Hussein. Sceptical but intrigued, I performed an experiment to prove the opposite. The result clearly demonstrates absurdity and futility of this bold assumption. Observe.

PHP: array_merge versus array union operator (+)

Seva — Thu, 22 Jul 2010 09:10:52 +0000

It comes that even experienced PHP developers don’t remember what is the difference between running array_merge((array)$a1, (array)$a2) and (array)$a1+(array)$a2. The answer is given @ php.net, no need to guess and argue.

Here is the difference:

Array Union: The + operator appends elements of remaining keys from the right handed array to the left handed, whereas duplicated keys are NOT overwritten.

array_merge(): If the input arrays have the same string keys, then the later value for that key will overwrite the previous one. If, however, the arrays contain numeric keys, the later value will not overwrite the original value, but will be appended. If all of the arrays contain only numeric keys, the resulting array is given incrementing keys starting from zero.

Phing plugin for Eclipse PDT

Seva — Wed, 14 Apr 2010 22:31:09 +0000

I love Ant integration into Eclipse JDT – it provides smart editor, handy auto-completion, and the most important – fully functional debugger.

Recently I have been laboring on porting a deployment system from shell scripts to Phing, a loose PHP port of Ant. And naturally, I miss the above. I still get a little aid from Eclipse – since Phing’s syntax is very close to Ant’s, I can at use Ant editor for Phing files to enjoy property navigation and target integrity validation.

I would be more than happy to announce that I’m going to fill the gap and implement Phing plugin for Eclipse PDT, but unfortunately – I’m too busy and too lazy. On the other hand, if you, my dear friend, will suddenly decide to accept this challenge, I can gladly invest my time in architecture, design, review & testing free of charge. Or should I anyway try to start it myself?

Heart-Touching Quotation

Seva — Wed, 14 Apr 2010 22:06:29 +0000

…When I was going to school we were always taught, “In the olden days of computing, computers were expensive and programmers were cheap. Now it’s the reverse. Therefore…” We are back to the future. At internet scale, programmers are (sometimes) cheap compared to the cost of electricity.

Kent Beck

Competing design attributes: performance vs. maintainability

Seva — Mon, 15 Mar 2010 22:51:42 +0000

Q. Is performance more important than other attributes like ease of use, maintainability etc? When designing your new code, what level of importance would you give to the following attributes?

A. You should not compromise on architectural principles for just performance. You should make effort to write architecturally sound programs as opposed to writing only fast programs. If your architecture is sound enough then it would allow your program not only to scale better but also allows it to be optimized for performance if it is not fast enough. If you write applications with poor architecture but performs well for the current requirements, what will happen if the requirements grow and your architecture is not flexible enough to extend and creates a maintenance nightmare where fixing a code in one area would break your code in another area. This will cause your application to be re-written. So you should think about extendability (i.e. ability to evolve with additional requirements), maintainability, ease of use, performance and scalability (i.e. ability to run in multiple servers or machines) during the design phase. List all possible design alternatives and pick the one which is conducive to sound design architecturally (i.e. scalable, easy to use, maintain and extend) and will allow it to be optimized later if not fast enough. You can build a vertical slice first to validate the above mentioned design attributes.

Is Facebook PHP HipHop bullshit?

Seva — Wed, 03 Feb 2010 00:05:49 +0000

Yes, it finally happened – Facebook announced the PHP to C++ translator. Hurray!
So, after reading the post I can assume it:

parses PHP code,
extracts PHP AST,
converts to C++ AST and
generates C++ code.

Honestly, it’s not the best way of optimization, but probably is a quick win (BTW, why did it take for them so long?), which theoretically might indeed generate efficient C++ code.

But wait, what do they report? “We’ve reduced the CPU usage on our Web servers on average by about¹ fifty percent³, depending on the page².”

I have some doubts:

Where is the objective and comparative statistics?
Did they even collect the general statistics, or just tested some pages to see the difference?
50%? That’s it? And how about the fact that C++ is actually 3-500 times faster (also here) (That would be 70% to 99% reduce of CPU). In fact the well known PHP Accelerators provide the same level of performance improvement as HipHop! It’s interesting, did they try any of them before entering the so called “Hackathon”.

In addition to reflections above, looking at the profiles of 2.5 coders which hacked the translator, I sorrowfully tend to estimate that the answer for the question in this Post’s subject is positive. Nevertheless I’m full of sincere hope that further info will dispel my suspects and prove that HipHop is actually a brilliant peace of software.

FreeBSD MySQL MyISAM multi-CPU concurrency performance

Seva — Tue, 02 Feb 2010 16:03:24 +0000

Recently we had a performance issue with multiple plain concurrent requests on a MyISAM table – the process list reported that all selects get stuck for very long time while Sending Data. After many tests we discovered that:

The problem only occurs when running on multiple CPUs;
The problem only occurs when running on FreeBSD OS.

After searching over the Internet we found, that this class of problems is being reported since antique versions of FreeBSD and, despite the claims it’s fixed with each version, it persists even in fresh 8.0.

So the decision was to install Linux on all the DB servers.

There is an alternative option though – it’s possible on FreeBSD to install MySQL of any version with WITH_LINUXTHREADS=yes parameter and it will be benefited from the Linux thread model.