Title: The Wall

Dear Francis

First, may I apologise for my English. I was born in Poplar, London's East End (yep, a real cockney) and never got to grips with the written word.

I started this mail half way through all the discussions on accu-general regarding reviewers experience and background, so I thought I would give you a quick breakdown in case you publish this so readers will know where I am coming from.

To summarise, April 2001 will be my tenth anniversary in the software industry, roles have gone from junior programmer through analyst programmer, senior engineer, technical/development team leader and consultant. The names given to roles vary by company but you get the drift. I started moving into management but maintained dirty hands, however this started to get too much with large teams. I am quite happy being responsible for 18 people but you just cannot find the time for good old hacking (not cracking, if you are confused see the new hackers dictionary). Languages range from assembler to point & click and wow out pops a ready to run (supposedly!) application. Most of my work has been back end C and UNIX for telecomm and network application companies. That should give you a general idea, you can have more details if you really want to.

If you remember any contact we have had in the past you may have noticed that I am a fussy person who likes detailed information. To give an example, I have asked for information about standards in the past and have been directed towards the likes of Plauger. Over the last couple of years (and especially since I started contracting in November) I am finding myself in positions where I am being asked to prove my statements relating to ANSI/ISO conformance.

There is a personal side to this story and a professional one.

For my own personal satisfaction I have decided to buy the necessary standards from ANSI/ISO (more on this later), the reason is that while there are some excellent books to use as a reference it is still the authors personal interpretation of the particular standard. Let me make it clear right now, that this is an issue of how a person thinks about questions, answers and information in general. I may never know as much about C as someone like Plauger does (I would love to!) but I am the type of person who needs sources of information, in this case the standards themselves.

From a professional point of view, I have argued with the senior language experts in several companies I have worked for. Many of these write the company coding standards and several are WRONG! It seems that if someone works for a particular company, working through the ranks for several years, their opinions go unchallenged. In the past, I have pointed out errors in company standards, first to the author and then to senior management. When the author has taken the time to listen to me (not very often) they get a confused look over the face and disappear to a meeting; never to be heard of again.

Though not acceptable, you may expect this from a new company just starting out, however I have seen this in more than one well-established multinational. The builds produce several thousand warnings, but hey a warning is not an error and we do not have the resources to check them all - so its OK to ship? Wonder why the bug reports run into thousands? Yes - warnings could be considered OK by some people in certain circumstances, personally I do not like them.

Back to the main reason for this letter. What I would like to see is an information centre where members can find information to learn or expand a particular skill. ACCU already has much of this information, particularly with the book reviews. What I would like however is something more condensed and focused.

Please note these are only ideas at the moment, on reflection and brainstorming with other members they may prove to be poor. As an example, we could have a section (sub-site?) broken down into software engineering (or just programming) topics, continue the refinement until we reach a certain point (i.e. programming, languages, C). Now in one place we first define the industry standards related to this topic, ANSI/ISO etc, along with de-facto standards; clearly stating which is which.

We could then have (and this could prove to be difficult) each of these topics (C in this case) broken into skill levels, i.e. introductory, intermediary, advanced, expert (not sure what to call them but would we need three or four levels; human nature says three as its a nice concept, market research says four, I think to stop people hanging around the middle). Anyway, at each level we could have as mentioned the standards first and foremost (I believe this is very important), we could then have a list (how many?) of recommended books and/or resources for someone at this level in this topic to learn from. After this we could have ancillary information, again related to this person at this level.

Wow, thinking about it this could take 20 years!

The point is I would be happy to take charge of this as it were. We would obviously need a great deal of discussion with many members and industry experts just to decide on a limited book list for each level. Its something I have thought of for some time, a lot of the information is out there; we need hunter/gatherers to bring it in, access it then display it for ever more. Well, what do you think?

Mark Batty

When it comes to programming there are certainly three kinds of standard that need consideration: language standards, coding guidelines and code presentation. The first of these should be provided by a copy of the relevant de jure standard (assuming that one exists), the last should not cause the ire that it does (the way I prepare code for publication is different from the way that I write code, monitors and paper provide different constraints). The middle one is where more consideration needs to be given. They should, generally, be guidelines and so capable of being overridden in some circumstances. However, a guideline that is in direct conflict with a language standard should be corrected at once. Anyone who tries to defend such a conflict is being less than professional.

Now this, the original letter and this issue's View from the Chair should start a lively debate. However, remember that words are cheap. FG.