$q = “SELECT `t`.`f` FROM `t` WHERE 1 “;
//populating sql filter
if (isset($filter) and is_array($filter)) {

foreach ($filter as $key=>$value) {
$q .= sprintf(” AND `%s` = ‘%s’”,$key, $value);
}

}

In this example i expect that filter should be the “equal” for all attributes, but this topic not for that.
Nicest thing is “WHERE 1″ in the first line: it allows you to not worry for content of coming filter, is there exists any useful (valid) parameter or not, and of course no cutting strings function in the end of populating (i like this point more).
in other words you say- “if i have no filters, than i have one … which is always true”, and i keep that one above others …
thanks

I can understand the architects and product leads who wants to have full documents and artifacts, the reasons are very strong, when everything that doing product managers, developers, and architects is written down and it is written in predefined templates, it is very ease to track everything considering project developing and product future, but in other hand first requirements of any product are budget, deadline, and quality, and then only strongly documented, but as shows experience technical and none technical documenting is may be the most expensive cycle in whole development process… so where is the Compromise…

I’ll try to explain some points of my opinion on development process model compromise

I think first of all well and fully documented should be only that products who already have got his place under the Sun, of course here some PM can be blame on me, but … “selavi” not every project that was born in someone brain and started development has at least succeed finished implementation.

second: there should be at least two different models described, one for complete new product second for new releases, i don’t think that it is good idea when you can pass some points of the model or start with the middle of described model according to your needs.

The man who suggesting concrete model should be responsible for success or fail of it, means if he project was failed there can be only fail of planing, good plan- plan to do, in that case model suggester will be more flexible while thinking about selecting model for concrete product.

It is not secret that the heaviest part of whole development process is planing and creating product skeleton, so i think it is better to pay more attention on that points, even in very first steps of product development process discuss more about which modules and technologies should be used product to get functional requirement that needed.

On the first pages of product documentation (vision …) should be reports of every team participants, product managers, marketing department, architects, developers …, other external modules delegates if exists…,

On the process of implementing should not be any big documentation in tasks of developers, but should be an strong requirement that every peace of code should be well commented by them, supervisors and architects should care about revisions of comments also.

of course the topic is larger then i can put in post, and topic itself is very interesting so i guess we will have hard discussions about it later on…

so limit of auto creating apaches by default is 256, we did decrease it up to 36. and decreased also some other potentials…

12 MaxKeepAliveRequests 15
13 KeepAliveTimeout 4
14
15 MinSpareServers 5
16 MaxSpareServers 12
17 StartServers 12
18
19 ServerLimit 64
20 MaxClients 32
21 MaxRequestsPerChild 16

for development it not big hurt, but it keeps your server up and running. And one note short command to see how many instances you have (this note i writing especially for me)

ps aux | grep apache| grep -v grep | wc -l

thanks a lot to Hovik and Arman for support