Timothy Cheng

Jatuh air mata depan satu dewan bila tengok best friend dah berumahtangga. Bukan senang nak dapat kawan yang ada masa susah dan senang I’m so happy for both of you. Please take good care of your wife Mey

What's the best Tennis Interview clip ever ?

Thailand is considering shortening the tourist visa from 60 days to 30 days due to widespread frustration with tourist behaviour. Main complaints by nationality by Thais: Chinese tourists: Travel in large, loud groups, often rude and disrespectful. Refuse Thai food, only eat Chinese cuisine. Frequent littering and ignoring religious site rules. Indian tourists: Large groups that litter beaches. Loud public behaviour. Aggressive haggling. Refuse to follow venue rules. Dominant presence in sex tourism areas like Pattaya. Low spending compared to other tourists. Many working illegally. Hygiene issues. Russian tourists: Drunken, disruptive behaviour. Shirtless in public. Fighting and rowdy antics Sense of entitlement. Working illegally. Israeli tourists: Strong sense of entitlement. Unpaid bills. Public intoxication. Aggressive and disrespectful behaviour. British tourists: Stag and hen night culture. Blackout drinking. Shirtless in public. Bar fights and disorderly conduct.

I personally have never been a fan of traffic cam feeds, compared to a precise ETA forecast based on good data. However, many people like verifying data with real ground observations—so I wanted to try to do something with the traffic camera feeds from Lembaga Lebuhraya Malaysia (LLM). Although the screenshot of traffic cams looks cool, if you actually look closely, the way LLM presents the pictures is almost IMPOSSIBLE to decipher. There are 5 main problems: 1) Flip-flopping: They mix cameras facing north and facing south, so users might actually be looking at the wrong side of the road without realising. 2) Info overload: They show lots of cameras which are actually close together in real life, just adding noise without any signal. For example, they have 7 cameras from the Menora tunnel in Perak—but just 1 would suffice! 3) Poor arrangement: The cameras aren’t perfectly ordered as you would see them along the road, so it’s hard to stitch a route in your head. The small labels make it even worse—need to squint if you want to figure it out (if you even can). 4) Irrelevant pics: Some images are not meaningful, e.g. the ones at rest stops don’t tell you anything about traffic. In short, more noise. 5) Broken feeds: Some cameras are offline, but they don’t remove the image, so a stale image is shown (note the daytime pic from Sungai Buloh when everything else is night). All of this makes the LLM user interface (as well as all the 3rd-party apps which just scrape the images and dump them on users) super difficult to use, imo. But, this problem can be solved with some data hygiene...