To improve code readability and ensure compatibility between modules from different developers, we strongly recommend following these naming conventions.
Root scripts and HTML templates — lowercase, words separated by underscores:
index.php
page_catalog.php
main.htm
page_404.htm
item_product.htm
Module scripts follow a three-part naming scheme, separated by underscores:
{company}_{group}_{purpose}.php
melbis. For your own
projects, choose a unique identifier: it will distinguish your modules
from the built-in ones and from third-party modules if you integrate
external code.Examples:
melbis_base_page.php — base page router
melbis_store_card.php — product card
melbis_cataloge.php — catalog
melbis_inc_logic.php — business logic library
melbis_block_slider.php — slider
melbis_client_auth.php — client authentication
Library names have four parts: the word inc takes the
place of the group, and the library’s own group moves to the next part
of the name — more on this below, in the “Module Groups” section.
In addition to the code itself, each module has parameters — cache, input data, table list, connected libraries. These are configured in the “Workbench” window, in the right panel next to the code editor, and are stored together with the module. This is described in more detail in the “Module Scripts” section.
The group in a module’s name carries semantic meaning and determines
how the module is displayed in the “Workbench” tree. Only one group name
is reserved — inc (also allowed:
include or an empty value). Modules with this group are
libraries: they cannot have HTML templates, and the parser does not call
them directly — their functions are used by other modules.
All other group names (base, store,
cataloge, basket, client,
block, cron, and any others) are defined by
you based on your project’s architecture. A group is simply a way to
organize modules in the file tree.
melbis_base_page.php — regular module (group base)
melbis_store_card.php — regular module (group store)
melbis_block_slider.php — regular module (group block)
For libraries, the word inc takes the place of the
group, so the library’s own group moves to the next part of the
name:
{company}_inc_{group}_{purpose}.php
In the “Workbench” tree, libraries are organized by this group —
alongside regular modules at the same level, but with their own icon.
There is no longer a shared inc folder where all project
libraries would be dumped:
melbis
auth melbis_inc_auth.php
logic melbis_inc_logic.php
web melbis_inc_web_callback.php
melbis_inc_web_topic.php
base melbis_base_footer.php
melbis_base_head.php
melbis_base_header.php
melbis_base_page.php
basket melbis_basket.php
cataloge melbis_cataloge.php
melbis_cataloge_sub.php
The purpose can be omitted if there is only one library in the group:
melbis_inc_logic.php represents group logic
with no purpose, while melbis_inc_web_callback.php
represents group web with purpose
callback.
The idea is to keep a library close to the modules that use it:
melbis_inc_web_callback and
melbis_inc_web_topic belong to the web part of the project
and are grouped under web, rather than being scattered
among all other libraries.
Each regular module contains a main function — its name exactly matches the module’s filename written in uppercase. This is the function the parser calls:
// File: melbis_store_card.php
function MELBIS_STORE_CARD($mVars)
{
...
}The exception is library modules in the inc group. The
parser does not call them, so they have no main function. Instead, they
contain a set of helper functions that are explicitly called from other
modules.
Helper functions within the same module are named with the main function’s name as a prefix, and their own name is written in PascalCase — every word capitalized, no underscores:
function MELBIS_STORE_CARD_Price($mTpl, $mId)
{
...
}
function MELBIS_STORE_CARD_Features($mTpl, $mId)
{
...
}The change of case marks the name boundary: to the left of it is the
module, to the right the function. The boundary is visible both to the
eye and to tools — the IDE and the linter split the name without
consulting the module tree, and a search for Price finds
the declaration as well as the short calls through
THIS()/THAT() (see “Utility Methods”). The
entry point tells itself apart from the helpers for free along the way:
it has no Pascal tail.
Abbreviations in the tail are written as words: SmsUrl,
XmlLoad — not SMSUrl, otherwise the prefix
boundary is lost.
This approach guarantees unique function names in PHP’s global namespace and makes it immediately clear which module a given function belongs to.
Functions in library modules (inc)
follow the same naming scheme — prefixed with the library name:
// File: melbis_inc_logic.php
function MELBIS_INC_LOGIC_OrderCreate(...) { ... }
function MELBIS_INC_LOGIC_OrderCalc(...) { ... }
function MELBIS_INC_LOGIC_OrderEdit(...) { ... }Projects with lowercase tails (
MELBIS_INC_LOGIC_order_create) keep working as before — PHP does not distinguish case in function names. The Pascal tail is a convention for new code.
The prefix in names exists precisely because all PHP functions lie in
a common space. A module or a library that has declared a space of its
own gets rid of it: the file name moves into the namespace
line, and only the Pascal tail is left in the function names.
// File: melbis_store_card.php
namespace MELBIS_STORE_CARD;
function Main($mVars) { ... }
function Price($mTpl, $mId) { ... }
function Features($mTpl, $mId) { ... }The main function in such a file is called Main: the
module name is already written above, and there is no point repeating
it. Its case is now the same as the helpers’ — what tells it apart is
not the spelling but the name itself, reserved for the entry point. The
parser looks for both forms, so flat modules with a main function named
after the file keep working.
The full name of a function changes predictably along with this: the
underscore before the tail becomes a backslash —
MELBIS_STORE_CARD_Price turns into
MELBIS_STORE_CARD\Price. The boundary between the module
and the function stops being a convention and becomes a language
construct, while a search for Price still finds both the
declaration and the calls.
For more on converting a module and on what to check while doing it, see the “Modular Scripts” section.
| Variable type | Rule | Example |
|---|---|---|
| Function input parameters | Start with $m, followed by camelCase |
$mVars, $mTopicId, $mTpl |
| Local variables | Lowercase, words separated by _ |
$id, $item_count,
$command |
| PHP constants | Uppercase, words separated by _ |
MELBIS_CACHE, MELBIS_LANG |
In a module’s main function, the input parameter array is conventionally named
$mVars.
A table alias is an abbreviation made of the first letters of
the words in its name (the {DBNICK}_ prefix does
not count): {DBNICK}_user_task → ut,
{DBNICK}_user_group → ug,
{DBNICK}_user → u. If a table appears in the
query more than once, or a role marker is needed, a suffix is added to
the abbreviation through an underscore: u_author,
u_exec, ug_main.
The rule is not cosmetic: the field hints in the Workbench are built
on it. The IDE reduces a typed ut. or
u_author. to the tables that fit the abbreviation and shows
their fields.
$command = "SELECT ut.id, ut.name, ut.state_key,
u_author.login AS author, u_exec.login AS executor
FROM {DBNICK}_user_task ut
LEFT JOIN {DBNICK}_user u_author
ON u_author.id = ut.user_id
LEFT JOIN {DBNICK}_user u_exec
ON u_exec.id = ut.exec_id
WHERE ut.state_key <> 'kClose'
";Values are not glued into the query text: the SQL holds a
:NAME placeholder in uppercase, and the value goes into the
parameter array — see the “Working with the Database” section for
details.
All template engine keys are written in uppercase, with words separated by underscores:
{TITLE}
{PAGE_NAME}
{PRODUCT:PRICE_OLD}
{#ITEMS}
<a href="?id={ID}">{NAME|html}</a>
{ITEMS#}