Applying the Façade Pattern to Government Services and Public Complaints

This article is written with the aid of the free version of Google Gemini AI 3.1 Flash-Lite standard mode As seasoned professional software engineers, you probably all know what the Façade Pattern is, and it’s so natural for you that it’s likely one of your second natures. However, what would happen if we try to apply the wisdom behind this design pattern into real lives? Even myself is somehow surprised by what I’ve found, so let’s explore this together :) \ Analogy Of Good Vs Bad Facades Before the official announcement of GenAI, we’d normally first type precise keywords in search engines, then manually judge which search results match what we want, and finally explore those link one by one to extract all the relevant info. With GenAI beginning to mature, we just have to instead input accurate prompts, and it’ll automatically summarize our desired contents for us, although we’ll still have to verify and fine-tune some important details. For instance, when I’m writing this article, I inputted the following prompt into the free version of Google Gemini AI 3.5 Flash extended mode: Lists the total amount of natural resources directly and indirectly consumed per Hong Kong people every day And the following is its outputs: Hong Kong is famous for its hyper-dense urban efficiency, but because the city relies almost entirely on imports, its actual resource consumption—especially when you factor in the "hidden" or indirect costs—is massive. According to data from WWF Hong Kong and environmental research groups, if everyone on Earth lived the lifestyle of the average Hong Kong resident, humanity would require 4.4 Earths to sustain its resource needs. The table below breaks down exactly what an average Hong Konger consumes every single day, contrasting direct personal usage with the indirect footprints embedded in the global supply chain. Daily Per Capita Resource Consumption | Resource Category | Direct Daily Consumption (Visible) | Indirect / Virtual Daily Consumption (Hidden) | |----|----|----| | Water | 326 liters used via municipal tap water (showering, cooking, washing). | 4,727 to 4,782 liters embedded in the production and transport of imported food. | | Food & Solids | 1.15 kg of solid food actually eaten (with a heavily meat-intensive baseline). | High material footprint from livestock feed, global shipping logistics, and extensive packaging. | | Energy / Electricity | ~16.1 kWh of electricity shared through household use and local grid infrastructure. | Extensive energy footprint used to manufacture and fly/ship consumer goods into the territory. | | Waste & Carbon (Output) | 1.51 kg of municipal solid waste sent directly to landfills. | 5.7 kg of CO₂-equivalent generated daily just from the supply chain of the average citizen's diet. | Deep Dive into the Footprint 1. The Water Paradox: Direct vs. Indirect While Hong Kong has successfully lowered its direct municipal water abstraction down to 326 liters per day per person, its indirect water footprint is one of the highest in the world. Because Hong Kong consumes an exceptionally high volume of meat—averaging around 126 kg per person annually—thousands of liters of "virtual water" are consumed overseas to grow the crops and raise the livestock that end up on local plates. 2. Food and Physical Materials The Centre for Food Safety's population surveys show that the average adult directly ingests 1.15 kg of solid food and 1,741 ml of liquids daily. However, the indirect natural resources required to make this possible are heavily amplified by global food supply chains. Over 90% of Hong Kong's food is imported, meaning the physical footprint includes the topsoil depletion, fertilizer usage, and fuel burned to cross oceans. 3. Energy and Waste On the local grid, a high reliance on air conditioning and dense vertical infrastructure results in an estimated 16.1 kWh of electricity per person, per day . On the back end, this consumption generates roughly 1.51 kg of physical municipal waste daily per capita, alongside a massive global carbon footprint embedded in electronic imports (with Hong Kong tracking as one of the highest e-waste producers per capita in Asia). The Biocapacity Deficit: Hong Kong's internal natural resources (biocapacity) stand at a meager 0.01 to 0.03 global hectares (gha) per person. With an actual total footprint of 5.4 gha per person, the city runs a massive ecological deficit, meaning it relies almost entirely on borrowing the natural resources of other countries to sustain its daily velocity. With the following sources cited: https://www.wwf.org.hk/en/cities/footprint/ https://pmc.ncbi.nlm.nih.gov/articles/PMC6588220/#:\~:text=During the last years%2C the city of,in data availability on urban food consumption . https://pmc.ncbi.nlm.nih.gov/articles/PMC6588220/ https://www.eeb.gov.hk/food/download/committees/acfeh/doc/2021/paper20210629_03.pdf https://www.statista.com/chart/35292/e-waste-production-per-capita-asia/ https://www.wwf.org.hk/en/?12300/WWF-Figures-Reveal-Hong-Kongs-Ecological-Deficit-is-the-Largest-in-Asia#:\~:text=The latest figures for Hong Kong's Ecological,3.1 Earths to fulfil our resource needs . While there might be some important details being missed or wrongly interpreted(it’s even possible to have hallucinations I’m unaware of), and it’s far beyond my capability to verify, this results still correspond to a good façade for me – massively reducing my workload from what I can’t handle to a clean and neat summary of what I’m interested in. I don’t have to think of what keywords to type in search engine, I don’t have to pick which search results are relevant, and I don’t even have to manually read the contents within those links. All I’ve to do is to carefully input a prompt that will get what I want, so in this regard, GenAI can be a good façade of integrating wanted info via search engines. \ Now, let’s say I’ve instead inputted the same prompt into a bad GenAI being a bad façade, and it just shows the aforementioned links. While it’s still much better than having to use search engines manually, why do I have to use such a bad AI, when I’ve free version of Google Gemini AI 3.5 Flash extended mode, albeit with limited quota for every several hours? Needless to say, if any modern GenAI were that bad, it’d surely be soon out of business for the greater good. \ Government Façade of Queries and Complaints Now, let’s extend this analogy to government One-Stop Services(OSS) or Single Point of Contact(SPoC), and it can lead to very surprising results even for myself. For convenience, I’d summarize them as the façade of centralizing civilian queries and complaints towards the government, and I’d abbreviate this as GFQC ( Government Façade of Queries and Complaints ). It’s responsible for taking all those requests, forwarding them to respective departments, while tracking the processes and returning the end results for the civilians involved, to boost government effectiveness and efficiency. First, let’s show what a bad GFQC will do. It’ll just tell its clients what departments to submit those queries and complaints, how to submit them, and all the relevant policies involved. Of course, it’s still already much better than nothing, because at least the civilians don’t have to guess what to do when submitting their queries and complaints towards the government. If their submissions don’t work as expected, they can compare the difference between what the GFQC claimed and what actually happened, then ask for both sides for further clarifications and corrected directions. It’s like using bad GenAI just acting as even more advanced search engines, they’re still much better than using ordinary search engines, as bad GenAI will still pick all the relevant links for you, and inputting an accurate prompt there is already much easier than typing precise keywords in ordinary search engines. But when there are already good GenAI that will just summarize the relevant end results for you, “search-engine like” GenAI is still bad GenAI in comparison. The same largely applies to a bad GFQC , as it’s just essentially an advanced search engine for its clients. It’ll pick all the relevant departments, policies, and submission requirements for civilians, but they still have to manually submit their queries and complaints to the respective departments, and then actively track the processes all by themselves. \ This leads to the concept of a good GFQC – Civilians can directly submit all their queries and complaints towards the government via this GFQC , and it’ll actively notify its clients for the respective processes of those submissions. GFQC doesn’t even need a standardized form across all departments, it just needs to tell its clients what forms to fill, what documents to submit, and the fee demanded by the respective departments, then GFQC will follow the rest for its clients. If they need further assistance, GFQC can even teach them how to perform all those steps they’ve to take. Do note that the GFQC doesn’t need to know the internals of any department, it doesn’t even need to order them for anything, it just has to honestly record all their actions visible to itself. A good GFQC will then actively update its clients about the processes of their submissions via the submission forwarding table containing the following fields: 1. Date – It’s the date of the key event occurred 2. Currently Responsible Departments – It always starts from the GFQC itself, and whenever the submissions are successfully forwarded to other departments, the subsequent records of this field will be marked as such 3. Department Actions – It can be “Validating”, “Rejected”, “Forwarding To (specific departments)”, “Forwarded To (specific departments)”, “Investigating”, “Needing Supplementary Documents”, “Resolving”, “Resolved”, etc 4. Remarks – It can be expected validation time, rejection reasons, legal basis of forwarding, relevant performance pledge, involved public policies, document requirements, attempted corrective actions, end results, etc If needed, the GFQC can even attach relevant public documents as further backup. The following shows the Finite-state machine of the submissions of complaints: \ As that of queries are even simpler, I don’t think I’ve to show that as well :) All the fields in the submission forwarding table must be filled, more specifically: - Both Date and Currently Responsible Departments must be filled by GFQC itself - Department Actions must be provided by the currently responsible departments, if they refuse to provide anything, the GFQC must clearly state in this field that, those departments refused to provide anything - Remarks is more complicated and convoluted. Since it’s the GFQC who does the basic validation before forwarding the submissions to relevant departments, expected validation time must be filled by the GFQC itself; Similarly, if the submissions are rejected in this phase, the GFQC must also state the rejection reasons; If the submissions are forwarded by the GFQC itself, it must also state the legal basis of forwarding, but if those submissions are forwarded by the other departments, they must inform the GFQC about the legal basis behind, if they provide unreasonable legal basis or refuse to provide it at all, GFQC will clearly state this fact at the end of the remarks; The rest of the remarks basically follow similar logic for increased transparency for the clients. Let’s focus on the following hypothetical example of this submission forwarding table: | Date | Currently Responsible Departments | Department Actions | Remarks | |----|----|----|----| | 2Jan2026 | GFQC | Validating Complaints | Expected To Finish Validation By 4Jan2026 | | 3Jan2026 | GFQC | Forwarding To Department A | Cap I Section J As Legal Backup | | 5Jan2026 | Department A | Investigating Complaints | Performance Pledge Of Finishing Investigation Within 7 Working Days | | 10Jan2026 | Department A | Resolving Complaints | Performance Pledge Of Finishing Resolution Within 14 Working Days | | 17Jan2026 | Department A | Needing Supplementary Documents And Extra Fee | (Details of the needed documents and fee) | | 20Jan2026 | Department A | Resolved Complaints | (Details of the resolution statuses) | | 24Jan2026 | GFQC | Closed Complaints | Clients Confirmed Acceptance Of Resolution Statuses | It’s all good right? Unfortunately, even with such a good GFQC , things can still go wrong, as shown in this example: | Date | Currently Responsible Departments | Department Actions | Remarks | |----|----|----|----| | 1Mar2026 | GFQC | Validating Complaints | Expected To Finish Validation By 3Mar2026 | | 2Mar2026 | GFQC | Forwarding To Department A | Cap I Section J As Legal Backup | | 4Mar2026 | Department A | Investigating Complaints | Performance Pledge Of Finishing Investigation Within 7 Working Days | | 14Mar2026 | Department A | Forwarding To Department B | Cap K Section L As Incorrect Legal Backup Provided By Department A | | 17Mar2026 | Department A | Forwarding To Department C | Department A Refused To Provide Legal Backup | | 21Mar2026 | Department C | Rejected Complaints | (Reasons provided by department C) Marked As Unreasonable By GFQC | | 26Mar2026 | GFQC | Closed Complaints | Clients Started New Complaints Towards Department A and C | But at least, now civilians know exactly what’s going on, so as long as the government has to face the public opinion, it’ll have no choice but to improve its effectiveness and efficiency, or the ruling party would lose in the next election. \ Limitations Of Good GFQC Of course, to implement a façade, its abstraction level needs to be above the modules it hides, meaning that no module should ever be able to control that façade, or that module could use it to hijack other modules; Similarly, to establish a GFQC , it needs to be directly under the governing team, with the restrictions that no senior officials under any other department can ever work under GFQC , to prevent it from being hijacked by those departments. Then the governing team needs to personally update internal rules and regulations governing the GFQC itself, meaning that any department head needs to directly negotiate with the governing team to force the GFQC to side with that department. Also, the government needs to sustain judicial independence, and maintain both government and legislative democracy, with the key that the legislation must have the right to summon the head of the GFQC for direct interpellation. It’s to minimize the risk of the GFQC intentionally forging the facts of the submitted queries and complaints, only then the GFQC can really function as intended, or it’d be just a decorator for the underlying mess. Therefore, unless an authoritarian government is very confident on itself, it’d have no incentive to run a good GFQC . To further ensure the power of the legislation when overseeing the GFQC , the senate should have access to the aggregated table of the GFQC performance for each major type of queries and complaints, with the following fields: 1. Responsible Department 2. Number Of Successful Forwarding Attempts To These Departments 3. Number Of Failed Forwarding Attempts To These Departments 4. Number of Incorrect/Unspecified Legal Backup For Refusing To Forward To These Departments 5. Average Number Of Days Taken For Forwarding To These Departments 6. Number Of Successful Forwarding Attempts From These Departments 7. Number Of Failed Forwarding Attempts From These Departments 8. Number of Incorrect/Unspecified Legal Backup For Forwarding From These Departments 9. Average Number Of Days Taken For Forwarding From These Departments 10. Average Number Of Days Investigating Complaints 11. Number Of Complaints Rejected 12. Number Of Rejections With Unacceptable/Unspecified Reasons 13. Average Number Of Days Resolving Complaints 14. Number Of Complaints Resolved 15. Number Of Complaints Failed To Resolve 16. Number Of Resolution Failures With Unacceptable/Unspecified Reasons In general, the higher the numbers in the 2nd, 6th, and 14th field, the better the performance, whereas the higher the numbers in the other fields, the worse the performance. Also, the aggregated table for each major type of complaints should show the number of complaints and the sum of each field too, and the same logic applies to those of queries. Of course, if the number of complaints is small, but those in the 2nd and 6th fields are large, than it’s actually a problem. Lastly, the 15th field includes those claimed to be successfully resolved by the departments but denied by the civilians. The government can even mark these aggregated facts as confidential, so the senate can only access them when the head of the GFQC is under their public interpellation, meaning that they can’t disclose any of these to the public, unless they’ve explicit prior approval from the Supreme Court. Even with such restrictions, the opposition party can already supervise the government more effectively and efficiently, as they know what and how to question the government. Another risk is the Two-party System , where the 2 major parties are used to ruling party alternation, the opposition party will have great incentives to act like it’s pushing the government hard, but has actually made secret deal with it, so the real pains exposed by the GFQC will ironically be swept under the rug. Therefore, Multi-party system generally has a significantly higher chance of successfully implementing a good GFQC , when compared to the Two-party system. Finally, with the existence of a good GFQC , working in the government will be much more painful and stressful, so the society needs to significantly raise their benefits and statuses, otherwise either the internal oppositions would be too much for the GFQC to work at all, or the best candidates would sway away from government jobs like a plague. This means that, a poor country will be extremely hard to implement a good GFQC , even when the whole nation wants to. But in a well-developed country, a good GFQC can actually link the salaries of the department heads(including that of the GFQC itself) and even those of the governing team, to the performance revealed by the GFQC . As those submitting queries and complaints are the only ones having the rights of publicly disclosing their submission forwarding tables, and aggregating them among all those clients can reproduce the aggregated GFQC performance table, if GFQC ever tries to forge facts there, it’d risk being exposed by the opposing party. Therefore, the legislation can just give raises to department heads honestly performing well, freeze salaries of those honestly performing poorly, and give the dishonest ones pay cuts, while adjusting those of the governing team based on those of the department heads. \ Façade Of MMO Metrics To further demonstrate the value of a good GFQC , let’s move on to measuring the performance of MMO. Let’s assume that you’re responsible for maintaining a MMO that’s already in production, but you don’t have access of any hardware performance, like the loading of the server CPU, NPU, RAM, VRAM and network, as well as the database maintenance statuses and some other important metrics. Of course, it’d be almost impossible in reality, but how would you ensure effectiveness and efficiency of the MMO for millions of players, when you don’t even know what’s really going on? On the player side, if they don’t know the FPS and network statuses of the game, and don’t even have software benchmarking their CPU, NPU, RAM, VRAM and network performance, how can the players know exactly what’s wrong? They can clearly tell that the game’s constantly lagging, but it’s of not much use when it comes to solving problems. Whereas if you do have accesses to these metrics, but without a centralized dashboard displaying everything at once, it’d be like opening an app displaying CPU load, another app displaying NPU and VRAM performance, yet another app displaying memory allocation, and finally another app displaying network volume. While it still works, especially when you’ve multiple monitors, but why would you’ve to do this, when you can have a centralized dashboard as the facade? By now, it should be obvious that, a MMO without any metrics is like a government without even a bad GFQC ; A MMO with only scattered metrics is like a government with merely a bad GFQC ; A MMO with a centralized dashboard is like a government with a good GFQC , with the submission forwarding tables of civilians being the client-side metrics of a MMO, and “the aggregated table of the GFQC for each major type of queries and complaints” being the server-side metrics for each server(or data center in case of really big MMO). With the fact that even a good GFQC is technically easy to implement without demanding much resources, but many governments don’t even have a bad GFQC , the real barriers barring the popularizations of good GFQC are mainly political and psychological, rather than any hard bottleneck. Of course, it’s very hard for governments to have a meteoric rise on just about anything, so even a bad GFQC is already a good start, and some governments are indeed heading this way. If one day bad GFQC becomes the norm, it’s only a matter of time before some nations will implement good GFQC , thus gradually raising the baseline of government effectiveness and efficiency to a whole new level, while benefitting lots of civilians along the way. Actually, those working in governments should seriously think about this: When you’re off duty, you’re just civilians as well, and it’s only a matter a time before you’ve to submit queries and complaints towards the government, wouldn’t it be good for you to have a good GFQC that can help better answering your queries and resolving your complaints? \ Summary 1. A bad GenAI is like a bad façade effectively just being an even more advanced search engines, but it’s still much better than using an ordinary search engine, as a bad GenAI will at least filter all the relevant links for you 2. A good GenAI is like a good façade saving most of your work from using search engines at all 3. A bad GFQC (Government Façade of Queries and Complaints) just teaches civilians how to submit what queries and complaints to which departments, so the civilians still have to do all the rest all by themselves, but it’s still much better than nothing, as a bad GFQC at least saves you from having to guess which departments to find 4. A good GFQC should centralize all submissions of all queries and complaints towards the government from all civilians, as well as actively updating them about the submission progresses and department forwarding statuses 5. Bad MMO metric displays scatter each metric in different benchmark apps, causing you to need many monitors to show them all at once, but it’s still much better than nothing, or you’d end up with endless guesswork 6. Good MMO metric displays centralize every metric into a single dashboard, making everything crystal clear, this applies to both the client and server-side, just like good GFQC letting the senate check the aggregated table too \