calimac: (puzzle)
[personal profile] calimac
The xkcd webcomic is widely and rightfully praised, especially its computer jokes - this one got a huge chuckle out of me - but this recent one on "how to fix computer problems" does not fit my experience at all, although others have linked to it approvingly.

I don't think that people who understand computers - or anything else, for that matter - have any idea how all-encompassingly baffling they are to people who don't understand them. I find many aspects of computers baffling myself - I hate to break it to my friends in the industry, but when you tell me what you're doing on the job, I usually have very little idea what you're talking about - and perhaps that humility assists me when I'm helping my mother with a computer problem.

True enough that I've had problems of my own that a Google search on some terminology has shed some light on, but when I actually am able to solve my mother's difficulties for her, the algorithm goes nothing like the one depicted in the cartoon. It's usually one of two things:

1) I've had the same problem myself at some time in the past, so I can now cut through all the frustrations and useless trials I went through at the time, and give the benefit of the actual answer that I'd finally come up with.

2) The answer is buried in the verbiage, in places where it's impossible for the inexperienced user to tell what is blither that may be ignored and what is useful information that may be encoded in unexpected ways.

As an example of #1, my mother currently does not receive e-mail from me. (We only discovered this when she noted in a phone call that she'd not received an e-mail I'd promised to send, which I'd sent two days earlier. Tests subsequently confirmed the problem and proved that it was limited to my ISP.) It is only because I've had past experience with this that I was sure this was a block and not a server glitch, because a glitch generates a bounce message and blocks don't. And while a technician would look first for a block in her personal e-mail program (as indeed the poorly-informed technician who visited her did), it is because of past experience that I know that ISPs regularly block other ISPs because some spam came (or appeared to come) from it, without informing anybody of it - not their own staff, nor their customers, nor the supposedly transgressing ISP.

As an example of #2, at one time my mother reported that she could neither send nor receive e-mail. She would press the "send/receive" button and get an error message. She read this message to me over the phone. It contained lots of gobbledegook but it also contained the letters "SMTP", which is how I knew the initial problem, at least, was in outgoing not in incoming e-mail. Now, there is no way she could have been expected to know, or remember if told, that SMTP is the end-user sending mail protocol, or to have known to have chosen this bit of data from everything in the error message to search on, or to have understood the neepery explaining it if she had. I happen to know this. Why? Because I've had to sweat my way through problems setting up SMTP and POP user addresses in the primitive past.

So my first advice was to try hitting the "Receive only" command. Should she have known about its existence already? Well, yes, we'd used it before. But it's buried in a pull-down menu and not something she regularly used, so she'd forgotten about it. It worked, and the problem eventually turned out to be an indigestible attachment to a "Reply To" e-mail that the pop server had let in on the original incoming message - and I should have thought of looking in the outgoing mail queue, instead of just blaming a server problem, long before I did.

So it's not all that simple, and inexperienced end users are at sea in all kinds of ways the intelligentsia can't even imagine (there have been occasions when I've been asked, "Well, exactly what is it you don't understand?" and I have to answer, "I don't know: if I did, I would understand it much better than I do"), and the real question is this:

Why, with all the user-friendliness supposedly built into computer programs nowadays, can't Outlook Express give an error message reading, "Sorry, I'm having trouble processing a message that you're sending" instead of giving a load of technical neepery with the letters "SMTP" buried in it?

Date: 2009-08-25 03:10 pm (UTC)
From: [identity profile] emerdavid.livejournal.com
Unhelpful error messages are not just a problem for non-techie people. I work with computer stuff all day, and I continue to get furious with software that produces messages that are so much gobbledegook, and could easily have been translated into something that actually tells you what the problem is. Even when they're not actually gobbledegook, the typically omit the crucial information you need, forcing you to set traces, inspect data dumps, etc. so that you can finally find out simple things like which database was unavailable, or which program had a logic error.

I have complained to IBM and other major software vendors about this, but the problem is endemic to the industry and no one seems to want to correct it.

Date: 2009-08-25 03:41 pm (UTC)
From: [identity profile] kalimac.livejournal.com
The thing that most mystifies me is when a program gives you an error code with a number (or some alphanumeric ID), but nobody at the vendor has any idea what it signifies. At some point, some programmer must have decided that particular errors generate particular codes, but nobody seems to have kept a list.

So why did they bother creating the codes in the first place?

The vendor attitude is usually that the computers came up with the error messages by themselves. But one of the few things I learned in my computer programming class is that computers don't write their own error messages.

Date: 2009-08-25 03:58 pm (UTC)
From: [identity profile] irontongue.livejournal.com
The error codes are for developers to use in debugging, most likely. The extent to which error codes are understood by tech support depends on the product. I can tell you that Documentum support knew and understood what particular errors meant. The chances that you can find that info for Outlook: slim.

> why did nobody keep a list

It's not that simple.

It's often the case that multiple conditions exist that can result in a particular error. The code that actually throws errors is scattered throughout a huge code base, in the case of big programs, like anything MSFT produces.

Error messages are difficult to document and that documentation is tough to maintain unless you've got an automated system for producing the documentation. The Documentum server had something like three thousand potential error messages, and you bet Tech Pubs did not document them. We could not have kept up.

Date: 2009-08-25 04:08 pm (UTC)
From: [identity profile] kalimac.livejournal.com
I'm not talking about Outlook Express, but programs in general. I haven't had this particular problem in a while.

Your explanation for the failure boils down to: it's incompetence.

Date: 2009-08-25 04:19 pm (UTC)
From: [identity profile] irontongue.livejournal.com
The issues are considerably more complicated than that, encompassing market demands, priorities set within companies that emphasize fast releases rather than ease of use and maintenance, the inherent complexity of computer systems design, how software projects are managed, what engineers are expected to, the general paucity of user experience designers, and so on. Sure, there's some incompetence in various levels and with individuals, but systemic and process issues are the major driver. Summarizing all of the issues as "incompetence" is wildly oversimplifying.

Date: 2009-08-25 04:26 pm (UTC)
From: [identity profile] kalimac.livejournal.com
Well, there are various reasons for incompetence.

I've seen examples of this at work in my own profession. Sometimes, when the thing that I've been assured can't be done is at a level I can do something about, I've gone ahead and done it myself. I broke the back of the uncataloged book reserve queue at a public library that way.

Profile

calimac: (Default)
calimac

January 2026

S M T W T F S
     1 23
4 5678910
11121314151617
18192021222324
25262728293031

Style Credit

Expand Cut Tags

No cut tags
Page generated Jan. 6th, 2026 07:45 am
Powered by Dreamwidth Studios