Archive for the ‘Serious’ Category

One Phone To Rule Them All

Wednesday, May 5th, 2010

Yeah, ok, it’s a bit of a gimmicky title :)

With my recent escapades into VOIP though, it quickly became annoying for me to have a desk IP phone and a softphone on my laptop, as well as my normal mobile and landline. Well, asterisk can route the landline as I want, so I guess that doesn’t really count. But given that I use my mobile for almost all of my calls (it is the number I hand out), it’s still annoying to have to swap phones.

Enter Siphon (click for screenies), a VOIP client for the iPhone. It is a real client, not a proxy like Fring (took me a while to figure that out :/). This means if you have a local Asterisk setup, you can use the local IP and actually get a reasonable latency :D The call quality is perfect with Siphon too. And it plugs into the addressbook in the iPhone, even better. The only down side is that apple don’t allow background applications to actually run on the iPhone, so incoming VOIP calls do not get routed through unless the application is left open. Oh well, it’s still a cool app :D

I wonder if I will have problems remembering when I’m calling someone over VOIP and when I’m calling over the mobile network given they are both made with my mobile phone device now…I don’t really want to spend an hour on the phone to a landline thinking I’m talking over VOIP only to discover a bill for $60 ;)

More Mac Woes

Monday, May 3rd, 2010

Another glaring oversight of having a Mac, and one that I think should be exploited by those making parodies of the Mac v PC ads ;)

I have a mac, therefore, I can do “everything” out of the box…except resize an image. Or create an image. Or, edit an image in any way. There is no Mac equivalent to MS Paint. Not that paint is very good, but at least in a pinch you can save a screenshot to paint and then resize it. Sure, the Grab utility in Mac OS X is great, but it’s not very much good to me at full size. Ohhh, right, I’m supposed to buy Photoshop, because that’s the only well known program that PC users are going to recognize that is also on a Mac that can edit images. Or I Google a lot and find a bunch of new programs I’ve never heard of, none of which are endorsed by Apple.

So, anyway, I’ve settled on SeaShore, which is based on GIMP. I didn’t like GIMP because it runs in X11 and doesn’t interpret key commands “properly”. By properly, I mean the same way Cocoa apps behave. You have to click on a window before you can click on any buttons in X11 i.e. if you are looking at the image and want to change tools, you must click the tool twice. One click selects the tool window, and one selects the tool. Then you have to click twice in the image to start using the tool. Annoying. Oh, and all of the commands are mapped to control button, not the command button. Also annoying.

SeaShore is a very cut down version of the gimp, but it’s small, Mac friendly, and allows me to paste and resize my screenshot easily :)

I also stumbled across this blog entry from 2006 that discusses some of the free editing tools for Mac, all of which turned up in my search as well, but it’s easier to link there for a summary.

Macbook Pro Mic Location

Sunday, May 2nd, 2010

Recently I have started getting into VOIP servers, specifically, Asterisk based VOIP servers. I use a Macbook pro as my everyday work computer (clamshell mode while at the office).

Problem 1) The Macbook Pro has line level inputs not mic level inputs. So when I bought a medium-range Logitech Clearchat headset, it didn’t work.

Attempted solution 1: I have an external Toshiba Dynadock which has a sound card in it, and connects up via USB. The mac detects the sound card. The Sound preferences control panel sees the mic and responds to it. No other application responds to the mic. I’ve absolutely no idea why. The only problem I could find on Google that remotely resembled this was when people are trying to get sound into a sound recording application, and the recording input is set at the wrong sampling rate. I tried modifying this to no joy. I can even hear the mic if I put the input on “passthrough” mode with Rogue Amoeba’s LineIn application, but I simply can’t get it to work in an application. The most simple test I have is using Audacity and recording off that input. Sigh.

Attempted solution 2: I un-clamshelled my mac to use the internal mic. So I have the internal mic and internal speakers and I’m talking and I can hear them, and they can hear me, but they are hearing an echo of themselves. Why? Because the internal mic in a Macbook Pro is in the speaker grill, next to the speaker. So the other person talks, it comes out my speakers, and the mic picks up on some of it and routes it back to the person.

Why put a mic next to the speaker? Lots of people say “oh I used it to record my voice and it was ok”. Yes, but you’re not outputting through the speakers at the same time. Even the normal Macbook has the mic next to the camera. It makes SENSE to put it next to the camera, because when people do video calling, they are looking at (and therefore speaking to) the camera.

As much as I like Mac OS X, I don’t think I’ll buy another mac. I might buy a Vaio instead. Windows 7 is quite nice, and even if it does “go slow” after a year and require a reinstall, I have discovered that a lot of what I do is server-based anyway, meaning that as long as I have a core set of applications on whatever PC I am using, I don’t have to back up / reinstall much.

Back online - and faster now!

Tuesday, February 23rd, 2010

Ok, the house move is complete, the DSL is back online. And faster now too :) I don’t think I’ll bother shifting the site off to one of the public hosting servers now :D 24 / 1 at almost full speed. If Amnet are using Annex M I can probably unbalance that a little in favor of the upload throughput given the signal strength :)

USA Outsourcing

Monday, February 15th, 2010

I recently started using eLance for some work I am doing, and I wanted to see how others use it. I opened a couple of projects from the main page and found this one. A US company outsourcing some IT work, fine. I’ve seen some US freelances on the site ranging in bids from $20ph to $100ph. Sounds reasonable enough to me, a little like Australia. Then the asian / indian market freelances are operating at $12 - $25ph. This was also no surprise. This one job posting that I opened here though, $2 PER HOUR?? Isn’t that a bit low? I agree with a freemarket economy, but I also agree with rewarding good people (it helps with retention of those people / good business, which makes life easier in a freemarket). I was surprised anyone would try to get something done for that amount. Maybe they’re going for quantity not quality…

Spotlight / mds CPU Usage on Mac OS X

Tuesday, January 5th, 2010

I recently upgraded my 10.5.4 OS to 10.5.8. I’d been putting off updating the OS for a while since it’s one of the installations that I know always requires a restart, and I hate restarting. Despite the fact that my mac regularly does stupid things that necessitate restarting (protected memory is really useless against protecting the OS from application crashes if there is no protection from an application consuming all available memory for starters).

Now I consistently have to put up with mds / mdsworker consuming my entire CPU. Even when it’s not, it is often spamming my hard drive, and since that is the slowest component in the computer anyway, my computer slows to a crawl again. mds is the process for spotlight that crawls the hard drive indexing the files. Apparently this situation can happen with a corrupt Spotlight database. This wouldn’t surprise me, because with my 200G hard drive, the Spotlight DB grows to 18G and doesn’t stop. I run out of space then, so I remove the DB for / and it starts all over again.

Now, I’ve noticed that mds is actually run by launchd (a variation of cron). And launchd actually allows limits to be specified:

launchd limit cpu 100

Limits the CPU usage to 100%. This should, in theory, help. In practice, it’s hard to tell. I can’t tell how the limit is achieved. If it’s an average, it would permit a spike of 150% if followed by a 50% usage interval of the same time. Then the question becomes how long the calculation envelope is (the time interval for the average) to determine how long a 150% CPU usage (more than 1 core) spike actually is. I don’t have the information. I can see that sometimes launchd / mds is still using >100% and sometimes it isn’t, but it moves too quick for me to see if the average is hitting 100% or not. Oh well, was worth the try I guess.

I have a feeling that running my CPU at full load and my HDD constantly being railed, and for some reason my GFX is getting worked over as well, resulting in system loads consistently above 2 and permanent temperatures above 70 degrees celsius for the CPU and GFX, is going to result in my HDD dying soon. I have everything important backed up, but I certainly won’t be forking out $3000 for another Mac Book Pro if that happens. I haven’t had any problems with Windows 7, and if I do, it won’t be unexpected so I’m prepared for a reformat in that case. And it will cost be $2k less.

New Server and ISPCP Exporting

Friday, November 13th, 2009

Well, I finally got around to moving this site (and the other test sites) off the old K6-2 300 and into a VM on a Core 2 Duo. There should be less downtime now, as it’s easier to manage!

 

The main delay was getting the import / export procedure for a domain in ISPCP correct. I now have a Perl script that will successfully export the mail and web files, and the SQL data for a domain from an ISPCP setup to a portable format that can be reimported on the other end. Yay!

 

I have posted the files on the ISPCP Forum if anyone wants to take a look. Actually, I’ve already made some minor bug fixes that I noticed after I released those versions, but I’m not going to bother updating them unless someone gives some feedback. If no one gives feedback, I’ll assume no one actually used them, and I’m not going to waste effort messing around with talking to myself on the internet - I do that enough here already ;)

Good Speech / Presentation Principles

Sunday, November 8th, 2009

Back when I was finishing high school, the TEE English subject required the demonstration of oral delivery capabilities of students. This was true oral delivery, not presentations with power point, although there is much overlap in the principles. Our teacher at the time gave us a handout that summarised these good principles. The handout was taken from a book, but the handout doesn’t have any indication as to the book it was taken from (it doesn’t even have page numbers!), and Google doesn’t pickup the text fragments. I tried to find another summary out that included such a good overview, but was surprised when I couldn’t. I’m sure if I altered my keywords and moved off the first couple of pages I could, but I type in excess of 100wpm when I want to, so it’s probably quicker to retype this. Here it is.

Manner of Speech

Our manner of delivery or the way we use speech may also affect how the message we send is received - it can either enhance the effectiveness of our speech or inhibit it. Our manner of speech is affected by the following:

  • Tone, or the emotional message of our speech. This reflects our attitide to and feelings about the subject.
  • Pitch, or the highness or lowness of speech sounds. he pitch of our voice affects its pleasnt appeal and therefore the listener’s willingness and ability to pay attention.
  • Intonation, or pattern or variations of pitch in speech. By varying intonation patterns we give certain aspects of what we say greater importance over others.
  • Pace, or the speed at which a person speaks. Well-paced speech is easy to listen to. By varying the pace we can control the degree of excitement we wish to communicate.
  • Loudness, or volume of seech. By altering the volume of what we say we can manipulate the audience’s attention.
  • Stress, or emphasis is given to syllables or words in a sentence. This helps our listener to sort out what is the most important idea or the most relevant evidence.
  • Rhythm, or pattern of a group of stresses. Creating rhythm gives empasis to what is being said and helps the listener to remember what is said, especially if the rhythm is pleasant.
  • Length, or time speech sounds last. If the length of speech sounds are well modified they are easy to listen to. By varying them we can affect the level of stress certain word and phrases carry and this can alter the over all meaning of what we have to say.
  • Body language, or non-verbal language, such as facial expressions (e.g. smiling to show agreement; frowning to show disagreement), gesture (e.g. use of the hands to mock or question), posture (e.g. way of standing to show confidence; leaning forward to stress a point), eye contact (e.g. winking to convey approval). Non-verbal language helps to provide an emotional context for what we have to say so that the listener can interpret the emotional significance of our message.
  • Word choice affects our manner of speech and often reflects our knowledge of our intended audience. We can alter the level of language or degree of formality by varying our vocabulary and choice of words. Words may be emotive if we wish to persuade, factual if we wish to inform or argue, imaginative and personal if we wish to create, technical if we wish to address an audience in a specialist field. Word choice is closely linked to a speaker’s purpose. Careful word choice will allow the speaker to convey their message as effectively as possible.
  • Structure of sentences: Sentences affect the clarity and coherence of our ideas. The presentation of our ideas can be made more or less effective by our choice of sentence structure. Our depth of meaning, degree of detail and the emotional impact of what we have to say can be altered by changing the way words are arranged in a sentence.
  • Sequencing of ideas: The order in which ideas are presented in our formal speech often reflects their importance to us. They may flow from the most to the least important or build up to a climax, usually with a closing summary at the end. If ideas are ordered logically they will create a succussful line of argument rather than merely state an opinion.

Making A Speech

Most people are afraid of public speaking. However, this fear and nervousness can be overcome by thorough preparation, taking control of the situation and by practice.

Preparation

First you need to be clear about the purpose of your speech. Most speeches aim to inform, persuade or entertain. Often speeches combine purposes.

Second, you need to know your audience. Ask yourself:

  • Who are they?
  • What are their backgrounds, abilities and education level?
  • What do they know?
  • What don’t they know?
  • What do they want to know?
  • What is their mood?
  • What are their biases, prejudices and interests?
  • How do they feel about an issue?

Third, think about the content of the speech. Ask yourself what it is you want to communicate. Collect as much detail as possible and other related material including pictures, anecdotes and examples. Consider the topic carefully and get your facts straight. Sequence your ideas logically, in a step-by-step or date / time order. Write your main points on cue cards with a brief summary of supporting details. It will help if you find a central theme for your speec.

Structure

Speeches usually follow the structure of an introduction, body and conclusion.

Introduction

  • Address the audience appropriately.
  • Give the audience a clear understanding of what you are going to talk about and why.
  • Make your opening concise and arresting. This can be achieved by a question for the audience to think about, visual aids, impressive facts and statistics, a personal anecdote or story, or a joke.

Body

  • Provide the audience with important factual material. Don’t be afraid to remind them of what they already know.
  • State the main points and provide support from either evidence, a well-reasoned argument, or an appeal to authority such as a member of parliament or an expert in a related field (e.g. a leading genetics engineer if your speech is about an hereditary illness).
  • Order your points in a way which will suit your audience’s knowledge and interests.
  • Points must have a logical structure (e.g. organised from the least to the most important).
  • Restate the points to ensure your autdience will remember them.
  • Use brief notes as memory checks for your main points and supporting details. Summary notes are best written on cue cards and should only contain one point per card.

Conclusion

  • Your conclusion must be memorable. Make it so by summarising the main points of your speech and finishing strongly by using a forceful tone of voice, a call for action, a quotation or an anecdote. For example, the conclusion of a successful persuassive speech should leave the audience with the blief that that is indeed the only conclusion which can be drawn.

Delivery

  • Stand quietly before an audience and pause before starting your speech. This helps to create tension by letting them wait and anticipate what is going to happen. (Hint: Make sure you have checked the microphone before you begin to speak).
  • Be aware of your body language. About 90 percent of your confidence in yourself and your task will be conveyed by non-verbal communication. Try to project an image of confidence, competence and enthusiasm. Use your eyes - the windows of your souls - to establish contact with your audience. Try to make each individual feel important and recognised. Use your hands to convey enthusiasm. If your are nervous, grip something.
  • Speak directly and fluently. You should aim to sound energetic, enthusaistic, friendly, and sincere. Vary your voice to keep the audience’s interest: by dropping the volume you can force the audience to listen intently; by raising it you can renew their interest. Slow down to make a point: by varying your tone and pace you will help your audience distinguish between points. Use pauses for emphasis.
  • Vary your standing position. Lean forward to be close to the audience. Assert your authority by standing up straight.
  • Smile - it will help you an your audience feel comfortable.
  • Give your speech emotional appeal by using emotive words which will engage the audience in a personal and positive way. This will enable you to communicate with your audience at both the feeling and thinking level.
  • Use humour to lighten your speech and make it more entertaining.
  • Enliven your speech with human interest details. By relating anecdotes based on personal experience you can make them real to the audience.
  • Keep your speech short to avoid over-taxing the audience’s attention span and letting boredom set in.
  • Rehearse your speech two or three times before you deliver it. Practise in front of friends or family using the video or tape recorder. Keep a close check on your timing. (While walking or jogging you could pactise creative visualisations of yourself successfully performing before your audience).
  • Use your intuition if you want to do something spontaeous like telling a story or joke. If it feels right, do it.
  • Before giving your speech anticipate any questions you may be asked. Always be polite to the person asking the question. If you do not have the detailed information that is requested (such as statistics), agree to provide it at a later date. Feedack is important to help you develop skills and confidence. Choose someone that you can trust to give you an honest appraisal. Don’t listen to feedback that is not constructive. If someone asks you to provide feedback praise the speaker in a few areas and then suggest one area that needs improvement.

Firefox: It Makes Me Angry

Monday, September 21st, 2009

I started using Firefox before it was Firefox. I think it was Firebird. The main reason I started using it was the tabbed browsing. Then they introduced the ability to restore tabs on startup. So now I use tabs as short-term bookmarks. In a given work-session I might open up 30 to 40 tabs. Unfortunately, since Firefox 2, Firefox will evntually gobble up all my unclaimed RAM (usually about 2G) and an entire CPU core. The more tabs I leave open, the faster this happens. It results in a huge slow down in Firefox (5 seconds to open a new tab, 1 second delays in entering text before it avtually appears) and an eventual application crash. When it crashes it doesn’t save the tabs properly and I lose the most recent ones (about the most recent 10 or so).

Additionally, why do I have to have Google as my start page? No, I don’t want to use my current tab. No, I don’t wan to use all of my current tabs. No, I don’t want to use a bookmark. I want to use a blank page, because the whole reason that I open a tab without an address is so that I can type one in! It’s a stupid application. I hate it. I hate it marginally less than Safari (actually, I like Safari, but it is too basic for web development and provides no method for extending the capability with extensions).

SCP With No Encryption: Why Not?

Tuesday, August 4th, 2009

I’ve read lots of forum posts recently where someone asks how to turn encryption off for an SSH session; specifically for an SCP transfer. Every one I’ve seen has been flamed for asking this. One common response I see is:

“the encryption doesn’t take up enough of the CPU to warrant the kind of exposure on a modern CPU, you’re probably I/O limited anyway”

GARBAGE. On my Via C3 Nehmiah @ 900Mhz (it’s a 1.2Ghz chip in a 100Mhz FSB capable motherboard hence the slowed clock), my CPU taps out at 3.7MB/s on a 100Mb/s network. An it taps out on the SSH daemon, not the I/O time. Using NFS I can pull 9 to 10MB/s at CPU tap-out. Encrypting at wire speed DOES take up significant CPU time. Normal SSH terminal connections, sure, negligible. Bulk SCP connections, it’s real. Just take a look at the performance measurements taken on a Via C3 on this Linux / Via Padlock OpenSSH enabling tutorial.

Modern distributions of linux (i.e. kernel 2.6.27+ based), seem to have patched the OpenSSH (and hence the SSHD) to use the hardware encryption on the Via chip (Padlock), and I can pull 9-10MB/s at CPU tap-out on that with SCP. A P3 733Mhz also taps out at 3.5MB/s with the same Linux (Ubuntu) though, so it’s definitely the software being optimized for the Via chip.

At the end of the day though, on my local LAN (wired), I don’t really care about the encryption of the file transfer. What I care about is the ubiquity of the SSH protocol. I’m also the only one using it to access files, so I’m not using it to replace NFS, I’m just using it to access my private files (which are sometimes quite large) using the already-configured ACL (PAM). Why can’t I disable the encryption for the SSH data transfer in V2? Sigh.