Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:122995
Message-ID: <2eb57476-1952-44f7-8408-762faa4f5971@bastelstu.be>
Date: Sat, 6 Apr 2024 14:55:53 +0200
Precedence: bulk
MIME-Version: 1.0
Subject: Re: [PHP-DEV] [RFC] Casing of acronyms in class and method names
To: Juliette Reinders Folmer <php-internals_nospam@adviesenzo.nl>,
 internals@lists.php.net
References: <792b2282-b7a3-40dd-899c-daab55353316@bastelstu.be>
 <69107662-eb03-4b75-8eed-59dd2eed6559@gmail.com>
 <66103C7A.3040906@adviesenzo.nl>
 <80ac7108-e016-4ae8-87ad-50afa0c2f257@bastelstu.be>
 <37f1b0c6-a7c8-41f3-bf2f-7dd466bb31c9@app.fastmail.com>
 <66106DDF.6050308@adviesenzo.nl>
Content-Language: en-US
In-Reply-To: <66106DDF.6050308@adviesenzo.nl>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
From: tim@bastelstu.be (=?UTF-8?Q?Tim_D=C3=BCsterhus?=)

Hi

On 4/5/24 23:32, Juliette Reinders Folmer wrote:
> I also took the liberty to ask accessibility expert Nicolas Steenhout
> [1] for his opinion on this topic and he has given me permission to
> quote his reply:
> 
>   > From a human readability and an accessibility perspective, Camel Case
> are best when words are concatenated like that.
>   > The rule I’d use here is uppercase the first letter of a word. Then
> lowercase the others. Unless you are writing an acronym.
>   > So in your example, it should be `PerformHTTPRequest()`
>   > The underscore becomes a problem because if for some reason we’re
> writing code as an example and it gets underline for any reason, the
> underline gets lost
> 

Thank you. I must admit I find it a little hard to interpret the answer 
without also seeing the corresponding question, because the way the 
question was phrased might've influenced the answer.

While I personally disagree with the acronym casing, it is not too bad 
in the example identifier used ("Perform HTTP Request"). However the 
naming policy will not just need to make the average case look great, 
but also make the worst case look acceptable. If it cannot do so and 
needs exceptions, then it failed to be a useful policy.

Hoping it isn't too much of a request, would you mind asking Nicolas 
whether the answer changes, when facing the following extreme examples 
consisting of consecutive acronyms.

I'm intentionally writing them in their natural casing with spaces to 
hopefully not influence the response:

1.

PCG Oneseq 128 XSL RR 64

which for reference is "Permuted Congruential Generator One Sequence 
128-Bit state XorShift Low Bits Random Rotation with 64-Bit Output", 
with the abbreviations / acronyms matching the naming in the reference 
implementation / corresponding paper.

2.

PDO ODBC

which is "PHP Data Object Open Database Connectivity", but no one writes 
those acronyms out in full.

3.

XML HTTP Request

which is "eXtensible Markup Language HyperText Transfer Protocol 
Request". I'd argue that the name is not really descriptive in the first 
place, but it's another existing real-world example.

4.

UUID v4

Taken from Ben Ramsey’s UUID library: 
https://github.com/ramsey/uuid/tree/4.x/src/Rfc4122

5.

S3 Client

as in "Amazon Simple Storage Service Client". Taken from Flysystem: 
https://github.com/thephpleague/flysystem/blob/3.x/src/AwsS3V3/S3ClientStub.php

Best regards
Tim Düsterhus