BlackInkBottle's Ink - All Comments

re: Kopi Syntax

icelava — Wed, 02 Jul 2008 08:44:54 GMT

Is that an enum with FlagsAttribute so you can bitwise OR them together?

re: Kopi Syntax

Katon — Sat, 10 May 2008 14:18:33 GMT

Still deciphering.... Need more caffeine :-P.

re: Kopi Syntax

hannes — Fri, 09 May 2008 06:42:44 GMT

Looks like greek to me... even worse - kopi sounds like java :p

re: Value of C++, Or Do I Live in Denial?

sb — Sat, 05 Apr 2008 15:17:02 GMT

Woooo you better edit this above, I spoke too fast. D is nowhere near being interpreted, it will just be compiled and run in one go. By making your source executable and calling it with a shell, it gives the illusion of being interpreted. See -run option on the dmd compiler.

re: Value of C++, Or Do I Live in Denial?

sb — Sat, 05 Apr 2008 14:57:06 GMT

And just about D, I know you gonna love this...

Invariants and unit test can be integrated in the code as of 2.0. Unit tests can be run at compile time :) Sweet

http://mail.google.com/mail/#inbox/11909cc5168fb9b7

The more I look into D, the more I feel drawn to it. Guess what? I can also be interpreted, just like Python!

re: Value of C++, Or Do I Live in Denial?

sb — Sat, 05 Apr 2008 14:49:38 GMT

Nowadays, from what I read, they are 2 group of C++ users. Those that claims using "modern C++ techniques" and the other.

Using modern C++ techniques requires a good knowledge of the STL and Boost. Emphasis is on using meta-programming through genericity and using a combination of memory managers like auto_ptr, smart_ptr and weak_ptr.

These 2 things really are the advantages of C++ over all other languages: it has the most complete genericity system (especially with the coming C++0x) and you do not pay for what you don't need (eg. you have garbage collection only if you want it, using smart_ptr...)

Actually there is a language that has been written on these principles: combining the power of C++ without falling in the pitfalls of C++. It's called D. I'm waiting for an occasion to pick it up. It seems to be what I've been looking for for year. The power of C++ with the easiness of C#!

re: Using Trac to improve the efficiency of product development and project management

Tommy Skaue — Wed, 12 Dec 2007 09:15:50 GMT

Great guide!!

How about upgrade this guide to include installation of VisualSVN server (free installer from visualsvn.com). That would make a complete installguide for SVN+trac on windows with a great GUI (yes, Graphical) to control the repositories! :D

re: Get "My Location" on mobile phone without GPS, from Google Maps

sidyk — Wed, 12 Dec 2007 08:08:39 GMT

Nice post. heres my comments:

Pls note that such locality positioning is fully dependent on the telco's cellular infrastructure and telco's willingness to reveal their subscriber's mobile location and at an affordable pricing. Thats the major drawback...

nevertheless, Network dependent positioning have some good advantage over GPS:

- able to work almost just as good indoors, whereas GPS signals deteriorate to almost nil when you enter a building, underground or a road tunnel.

- easier to build more responsive LBS and other apps since locality data is available on the network; no need to fetch the coordinates from the client GPS mobile phone. (enabling Push rather than pull)

then again network-based flipside (disadvantages):

- accuracy is dependent on the density and capabilities of the telco's cellular topology. More dense the cellular base stations, the smaller the area of coverage of each base station hence more accurate the location. On the other hand, if the telco uses tall antennas to cover large radius of area, accuracy suffers

there are techniques such as Enhanced-observed Time Difference (EOTD), Time Difference of Arrival (TDOA), Angle of arrival (AOA) to help with the accuracy but nothing as close to GPS in terms of long-term reliablity.

theres more to this area of study, in fact there is a whole bunch of academic researchers in various universities working on it. Give it a few more years, and I believe the technologies will mature and LBS will truly flourish.....

re: Get "My Location" on mobile phone without GPS, from Google Maps

blackinkbottle — Fri, 30 Nov 2007 07:51:44 GMT

True, triangulation requires reading of signal strength from identifiable cell-towers and the exact coordinates of those cell-towers involved in the triangulation calculation. In fact only telcos know their assets and such info well.

However, I think added value of google's My Location feature is that it does not rely on being given such coordinates information from telcos (ok in fact this is pure guesswork :)). They probably relied on thousands of phones that DO have GPS installed. When such devices send information, both cell towers information and their CURRENT location is sent. When there are many phones sending from nearly everywhere, millions of surveying records could be accumulated and are being accumulated.

Algorithm could then be used work on such huge amount of records to deduce where the cell-towers POSSIBLY are. The database of cell-towers could grow in size and improve in quality over time.

This is the small step that I consider valuable. Of course, privacy might be a issue, but if it is claimed anonymous, I might have to believe it.

Somehow, the idea reaps the economy of scales of distributed computing, just in LTA's traffic condition system that constructs the loading information based on the info sent by all the roaming taxies.

re: Get "My Location" on mobile phone without GPS, from Google Maps

Softwaremaker — Fri, 30 Nov 2007 03:34:16 GMT

This is really nothing new - with the onset of Location-Based Services (LBS) in the late 90s.

I suspect that this is more a telco play more than the map-provider play. Telcos have to give permissions to their co-ords of their triangulated cell-towers so any provider can determine the geo-locations and decide how they want to provide value-add services to the end-consumers.

In fact, with the triangulation technology of GSM networks, it is well-known that privacy is an issue although telcos has given assurance for many years that they will respect it.

re: Using Trac to improve the efficiency of product development and project management

kgx — Mon, 29 Oct 2007 09:25:25 GMT

Thanks...awesome guide :-)

Another mind map on Integration Topology (Message Broker, Message Bus etc)

BlackInkBottle's Ink — Wed, 03 Oct 2007 03:24:29 GMT

Theearlier post of mind map focused on WCF. The below is more on integration topology, withmore contents

re: Readings and Questions on Software Architecture and Software Architect

blackinkbottle — Sat, 25 Aug 2007 05:03:16 GMT

Interesting posts on the developer ratio.

True, mentoring is the only way to bring up the level of the whole team and the quality of the final output. Brewing wine takes time, so hopefully the critical point happens sooner before any detrimental effect from underperforming staff (thinking of: rework on-behalf, too-frequent refactoring == rework, test team becomes unit testing team) happens on the project. That's growing pain: a lot of oversights and secrunity is necessary initially but hopefully only intially (this could turn tricky if some people feel intrusive this way but the outcome could become favorable, one case that the good end might justify the seemly "evil" means).

On the other hand, sometime we get chances to pick talents to assemble a team; but sometimes or most of time we would have to make use of existing resources. So working with constraint is the only way out.

re: Readings and Questions on Software Architecture and Software Architect

cruizer — Thu, 23 Aug 2007 01:18:21 GMT

i think an architect needs to code. otherwise, how would he/she know the problems/pain points the devs are encountering?

i like this blog entry by fred george. maybe a project overall needs more "masters" who can collectively act as architects...

http://processpeoplepods.blogspot.com/2007/08/ideal-development-skills-for-pods.html

re: Readings and Questions on Software Architecture and Software Architect

blackinkbottle — Wed, 22 Aug 2007 17:47:38 GMT

The problem with "need or need not software architect", most of time, is that we do not know how to measure his/her output realisticly and assess accurately the contribution.

The value of an software architect, in my opinion, is to inject valuable thinking (from past experience, and most likely bottom-up experience) in the design of software systems and by all means to convince others those all being said and painted is justifiable - and get responsible and liable for that. We are severely short of the latter. I believe defining concretely clear role, scope of work and accountable output is the key to separate the good, the bad and the ugly.

Communication is the key of an achitect. I relate this to the old debate, does architect need to code? My answer: on one side, why should an architect seem be ashamed of coding - does coding make an architec less worthwhile? (no!); on the other side, why should architect insist in the merit of coding if someone else in your team could do much better than you?! The ultimate goal of architect is to influence design, to communicate design, if coding is the best communication medium and means, then use it; if coding is standing in the way, then think of something else.