<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Lasan Mahaliyana on Medium]]></title>
        <description><![CDATA[Stories by Lasan Mahaliyana on Medium]]></description>
        <link>https://medium.com/@limahaliyana?source=rss-9020d25f0558------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*uEvdipXY0xRcz0NT</url>
            <title>Stories by Lasan Mahaliyana on Medium</title>
            <link>https://medium.com/@limahaliyana?source=rss-9020d25f0558------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 13:14:15 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@limahaliyana/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Quick Share: A comprehensive deep dive into AirDrop’s rival]]></title>
            <link>https://medium.com/@limahaliyana/quick-share-a-comprehensive-deep-dive-into-airdrops-rival-6a59c52ea999?source=rss-9020d25f0558------2</link>
            <guid isPermaLink="false">https://medium.com/p/6a59c52ea999</guid>
            <category><![CDATA[quick-share]]></category>
            <category><![CDATA[airdrop]]></category>
            <category><![CDATA[android]]></category>
            <dc:creator><![CDATA[Lasan Mahaliyana]]></dc:creator>
            <pubDate>Mon, 16 Feb 2026 19:56:57 GMT</pubDate>
            <atom:updated>2026-02-26T16:52:51.862Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*46Hu-LY8A_bHnNNkMCFDgQ.jpeg" /><figcaption>Quick share logo</figcaption></figure><p>Let’s face it. AirDrop is amazing. It works flawlessly to share files, photos, videos, whatever across the apple ecosystem. For years, us Android folk have been missing an alternative. I really hate writing unless I have something interesting to talk about and I figured this would be worth sharing with people. Let’s get into how it works.</p><p><em>This write-up was inspired by the open source git repo of</em><a href="https://github.com/google/nearby"><em> google’s nearby share</em></a><em>. I’ve been following the repo’s development and am currently developing a Linux implementation on </em><a href="https://github.com/kidfromjupiter/nearby"><em>my own fork</em></a><em>. So I’d say I know a few things about the darn thing.</em></p><h3>The Main Components</h3><p>Quick share (formerly known as Nearby Share) is primarily broken into 4 parts.</p><ul><li>Nearby Connections</li><li>Fast Pair</li><li>Nearby Presence</li><li>Nearby Sharing</li></ul><h4>Nearby Connections</h4><p>This is the brains of the operation( this sounds so corny like chatgpt wrote it ). But that’s the best way to explain it. All the other components are actually built on top of this. It handles all the wireless radios, encrypting the connections, advertising and discovering services, and the actual transmission of data packets (or frames. hmmm… I dunno how the OSI model relates to this. Who cares ¯\_(ツ)_/¯ )</p><h4>Nearby Sharing</h4><p>The star of the show (I just had the unsettling thought that either I’m starting to write like chatgpt or chatgpt is writing like me ). This is an abstraction layer over Nearby Connections. It tells Nearby Connections what kind of advertisements to send to other devices and what kind of advertisements to listen to ( we call this discovering ). It also handles identification of remote devices and some closed source magic around sharing with only your contacts. This is the closed source part is the source of many headaches for me.</p><h4>Fast pair and Nearby Presence</h4><p>Fast pair is the handy dandy popup you get when you open up a fresh pair of earbuds that we’re previously paired to your devices. I’ve seen fast pair in the repo and I know it works on top of Nearby Connections but apart from that I haven’t looked too far into it.</p><p>Nearby Presence is also built on top of Nearby Connections, and the only difference with Nearby Presence and Nearby Sharing is that it doesn’t form any long lasting connections with remote devices to share data. Only listens for other devices and occasionally screams “<strong>IM HERE” </strong>so that other devices can here it.</p><p>My work currently revolves around Nearby Connections and Nearby Sharing so we’ll not talk too much about these two.</p><h3>The Mediums</h3><p>The Nearby Connections library provides platform-level abstractions over several mediums ( this just means that other programs using Nearby Connections don’t have to worry about the platform specific implementations for the underlying radio hardware and such )</p><ul><li><strong>BLE</strong></li><li><strong>Bluetooth Classic</strong></li><li><strong>Wifi LAN</strong></li><li><strong>Wifi Direct</strong></li><li><strong>Wifi Aware</strong> (aka. Wifi NAN. Super cool tech. The standard isn’t new by any means but only recently getting widespread attention by the tech community )</li><li><strong>Wifi Hotspot</strong></li><li><strong>UWB</strong> ( Ultra Wide Band. Ohh yeah. This bad boy is supported as well. Although you’d have to have a pretty recent device to take advantage of this )</li><li><strong>WebRTC </strong>( Experimental at this stage from what I can see. But Android supports it I think)</li><li><strong>NFC</strong></li><li><strong>USB</strong></li><li><strong>AWDL</strong> ( Apples own proprietary thing. That’s what AirDrop works over )</li></ul><h3>How it works</h3><p>There are 3 main parts.</p><ul><li>Advertising</li><li>Discovery</li><li>Connection Initiation and Bandwidth Upgrade</li></ul><h4>Advertising</h4><p>This is basically saying “Hey, I’m a quick share capable device. You can share stuff with me”. There are multiple advertising mediums (we’re going to focus on the ones implemented on android here, as it represents the full capabilities of the library). You have BLE ( Bluetooth Low Energy ), Bluetooth classic and WiFi LAN. Each different advertising medium handled advertising in their own unique way. The library contains abstractions for interfacing with these different mediums.</p><p>Okay then. We advertise we’re capable of quick share. But there’s a LOT of devices that are quick share enabled these days. It wouldn’t be super helpful if whenever you wanted to share a file, every device around you popped up saying “Hi hi Hi Hi Hi hI hi hI hi hI Hi Hi”. So, every advertisements include some sort of metadata announcing what they’re device names are and exactly <strong>What</strong> sort of service they provide. Because what you have to know is that Nearby Sharing <strong>is an abstraction layer over nearby connections</strong>. And since nearby connections acts as a sort of framework for discovering and communicating with nearby devices, the nearby sharing abstraction puts a bunch of nearby sharing specific metadata into advertisements so that they we know which is which and who can share with whom.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YKJsx8czND8NOCNT_oWK2A.png" /><figcaption>this is what an advertisement frame looks like over BLE</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/918/1*lAhDVuwt0t5gd-kZOlbjtw.png" /></figure><p>If you direct your attention to the second image, you can see the 16 bit service class UUIDs and the service data fields for those respective Service IDs. The 0xfef3 service UUIDis specific to the Quick Share protocol. Anything else built on nearby connections may have other UUIDs. And the service data encodes all the metadata required ( Squint hard enough on the first image and you can make out Device-09487 . This was taken when I was testing my own implementation but if you’re actively advertising with quick share, you’ll see something like “Someone’s S23”.</p><p>And below is an example of how advertisements happen over wifi lan. You can see the _quickshare._tcp service name when I look for mDNS advertisements. ( Dude I swear to god the service name on wifi lan was an encoded mess the entire time I was working on this library and now its just cleartext? wtf happend? I go about my life for 5 days and this happens )</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/916/1*KgMWzd3Fn1X882aqLzA7pA.png" /></figure><h4>Discovering</h4><p>Since you now know about advertising, discovery is easy. Its just looking for these specific ‘tags’ in advertising packets (frames?) they stumble upon on the airwaves. For example, if you’re looking on BLE, you’ll look for 0xfef3 service UUIDs (that is if you’re looking for quick share devices. Different services may have different service IDs ) and try decoding the service data. ( There’s an agreed upon protocol for encoding the service data so everyone who’s Nearby Connection compatible knows how to decode it ).</p><h4>Forming a connection</h4><p>This is where it gets complicated. Each medium has it’s own method of forming connections. We’re going to skim over how each medium does it and focus on what happens when a connection is initiated. Basically, one device listens for incoming connections and one device initiates the connection. And depending on the nature of the medium, a decision is made to either encrypt the link or not. Encryption happens on inherently insecure links like Wifi Lan and Bluetooth Classic but isn’t encrypted on links like Wifi Direct or Wifi hotspot, where a threat actor would need a passkey to even join the network. (I’m no cybersecurity expert and you should take what I say about network security with a grain of salt. All I wanted was to share files between devices. Whether attackers want to see my university lecture notes and my half naked gym progress pics is entirely up to them )</p><p>Of course initiating a connection normally happens over an “open” medium like bluetooth classic, BLE, or Wifi lan. This is because you’d need to share credentials to form a Wifi Direct or Wifi Hotspot connection. Once a connection is formed over a shared medium, keys are exchanged using ukey2 , Google’s own take own Diffie-Hellman based authentication key exchange protocol.</p><p>Next, there’s two seperate paths this link could take. Either</p><ul><li>Link could remain as is</li><li>Link could be upgraded</li></ul><h4>Bandwidth Upgrading</h4><p>The library defines several tiers of link qualities. Wifi based mediums ( Wifi Lan, Hotspot, Wifi direct are high quality links. They support the highest bandwidth ) and Bluetooth based links are generally lower quality (BR/EDR, LE L2CAP )</p><p>After forming the initial connection, the library automatically exchanges information on what each device is capable of. Using that information, either one of the devices calculates their shared capabilities and sends a BANDWIDTH_UPGRADE_NEGOTIATION frame over their shared medium. They go back and forth and exchange a bunch of data and credentials, initiate the new connection and then close the old connection.</p><p>And voila! The bandwidth is upgraded.</p><p>So much trouble. So much work just to share a simple file. <strong>That’s the thing</strong>. It doesn’t have to be this way. That’s the promise of Wifi NAN ( aka Wifi Aware ). Wifi Aware <em>is</em> a supported medium of Nearby Connections BTW. Although I don’t think its supported as an advertising/discovering medium for some reason. On android, it’s used as an upgrade medium.</p><p>Wifi NAN supports discovering, advertising and sharing data through NDP ( Neighbour datapath protocol or something that rhymes with it ). This isn’t a new idea either. Wifi Direct had similar ambitions when it was first introduced. There are differences sure, but the fundamental operating principle is kind of the same. I’m not well-versed enough in either of these technologies to comment on this though.</p><p>There’s a lot I can’t say about this absolute behemoth of a software library without littering it with even more unintelligible technical jargon so I’ll leave this as is for now. Feel free to reach out if you want to know more about this.</p><p><em>Ciao</em></p><p><em>This article was also published on the </em><a href="https://blog.rotaractmora.org/quick-share-a-comprehensive-deep-dive-into-airdrops-rival/"><em>Rotaract Mora blog</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6a59c52ea999" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Modernizing the healthcare enterprise, A beginners guide for startups]]></title>
            <link>https://medium.com/@limahaliyana/modernizing-the-healthcare-enterprise-a-beginners-guide-for-startups-a9a4d223c648?source=rss-9020d25f0558------2</link>
            <guid isPermaLink="false">https://medium.com/p/a9a4d223c648</guid>
            <category><![CDATA[healthtech]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[healthcare-technology]]></category>
            <dc:creator><![CDATA[Lasan Mahaliyana]]></dc:creator>
            <pubDate>Sun, 09 Nov 2025 15:34:33 GMT</pubDate>
            <atom:updated>2025-11-09T15:34:33.512Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KIdTEGR9bAd_kp0sLXBGcA.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@owenbeard?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Owen Beard</a> on <a href="https://unsplash.com/photos/citiscan-result-hand-ok-DK8jXx1B-1c?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p>The healthcare industry is huge — literally. It concerns every one of us who has ever existed and ever will exist.</p><p>Over the years, the industry has undergone sweeping changes. The flow and ebb of human technology and scientific progress remain deeply intertwined with the fate of the healthcare sector. And yet, despite entering the digital age in the 1970s, why does it still feel like the healthcare industry is firmly rooted in the past?</p><p>Truth be told, healthcare is a different beast.</p><p>Surrounded by laws and regulations, the Silicon Valley “move fast, break things” mentality simply doesn’t work here <em>cough cough, Theranos</em>. A rough analogy: if we accidentally drop a database table in a social media startup, we’re just ruining someone’s cozy night-in binge-watching Netflix. In healthcare, we might quite possibly be killing a patient. There are real risks involved.</p><p>The good news? You don’t have to modernise healthcare — it’s already been done. All the interoperability standards and technologies have been developed, standardised, and are in use today.</p><p>The bad news? There’s a lot — like, a metric truckload of things you’ve got to know.</p><p>Since this is more of a blog article than a reference manual, we’ll only focus on the bare basics. Otherwise, we’d be here a while.</p><h3>Part 1: First things first. The standards, the words, the jargon</h3><ul><li>IHE</li><li>FHIR</li><li>HL7</li><li>DICOM</li><li>EHR</li><li>PACS</li><li>SNOMED CT</li></ul><p>Healthcare industry really loves their acronyms all right. Lets break them down. There’s a lot more, but this should be enough to get you going. DICOM, HL7, IHE, FHIR and PACS are especially important — if you’re planning to do anything patient facing that is.</p><h3>IHE: The playbook</h3><p>IHE stands for <strong>Integrating the Healthcare Enterprise</strong>. You can think of IHE like the rulebook for how different standards interact with each other. There are many different standards of information interchange in the healthcare industry. IHE ensures interoperability between the standards.</p><p><strong>Example scenario</strong></p><p>Say two hospitals want to share patient medical documents securely</p><p><strong>IHE XDS.b</strong></p><p>defines <em>how</em> to store and share clinical documents ( like discharge summaries or lab reports ) using standard formats ( FHIR, DICOM, etc)</p><p>Step Actor IHE Role Description 1 Hospital A <strong>Document Source</strong> Publishes a clinical summary as a CDA XML document to the repository. 2 Repository <strong>Document Repository</strong> Stores the document and sends metadata to a registry. 3 Registry <strong>Document Registry</strong> Indexes metadata (patient ID, author, date, document type). 4 Hospital B <strong>Document Consumer</strong> Searches registry for patient documents and retrieves them from the repository.</p><p><em>IHE defines the </em><strong><em>workflow</em></strong><em> and </em><strong><em>roles</em></strong><em>; it reuses HL7 CDA for document format.</em></p><h3>HL7: The language</h3><p>You may be thinking okay, so not JSON or XML or something simple? They just had to invent something of their own. Well no. Not exactly. Think of JSON or XML as the <em>medium</em> of language. When humans talk, its over sound. But there can be different languages. HL7 is kind of like that.</p><p><strong>Example scenario</strong> When a patient is admitted to the hospital, the Admission/Discharge/Transfer (ADT) system sends a message to the EHR to register the patient.</p><p>The following is an extract from a message that would be transmitted. The following is a message from the HL7 v2 standard.</p><pre>MSH|^~\&amp;|SENDING_APPLICATION|SENDING_FACILITY|RECEIVING_APPLICATION|RECEIVING_FACILITY|20110613083617||ADT^A01|934576120110613083617|P|2.3||||<br>EVN|A01|20110613083617|||<br>PID|1||135769||MOUSE^MICKEY^||19281118|M|||123 Main St.^^Lake Buena Vista^FL^32830||(407)939-1289^^^theMainMouse@disney.com|||||1719|99999999||||||||||||||||||||<br>PV1|1|O|||||^^^^^^^^|^^^^^^^^</pre><p><strong>MSH</strong> — Message header ( contains metadata ) <strong>EVN</strong> — Event type ( A01 means admit/visit notification)</p><p>You get the idea. I’m not going to list everything in the spec here.</p><h3>FHIR: HL7 on steriods</h3><p>FHIR was developed by HL7 to fit into the modern web standards.</p><p>Now dear reader, you may be thinking, “Hold on, you told us that HL7 is a standard. How can a standard develop another standard?!??”. Ahh.. yes. I lied. HL7 is also a standard <strong>and</strong> a standards development organisation.</p><p>It uses <strong>RESTful APIs</strong>, <strong>JSON/XML</strong>, and <strong>OAuth2</strong> — just like web apps and mobile services today.</p><p>The only thing you have to keep in mind at this stage is that everything is a <strong>Resource</strong> to the FHIR spec. Patient, Observation, Appointment, Medication, they&#39;re all a <strong>Resource</strong>. FHIR also names standardised endpoints such as /Patient/123 or /Observation?patient=123. Yes keen observer, it also supports URL Search parameters. How cool is that ?</p><p>HL7 v2 is still used in a lot of hospitals to this day by the way. But newer installations tend to focus more on FHIR.</p><p>Okay lets look at an example and get a better idea.</p><p>Lets GET some patient info. The patient ID will be 47958444. ( You can actually try this yourself. HAPI is a fully FHIR compliant backend coded in Java. They have some dummy data for testing )</p><pre>GET http://hapi.fhir.org/baseR4/Patient/47958444</pre><p>Here’s the response</p><pre>{<br>  &quot;resourceType&quot;: &quot;Patient&quot;,<br>  &quot;id&quot;: &quot;47958444&quot;,<br>  &quot;meta&quot;: {<br>    &quot;versionId&quot;: &quot;1&quot;,<br>    &quot;lastUpdated&quot;: &quot;2025-06-19T19:28:45.469+00:00&quot;,<br>    &quot;source&quot;: &quot;#VuociAlICXqhYDiI&quot;<br>  },<br>  &quot;text&quot;: {<br>    &quot;status&quot;: &quot;generated&quot;,<br>    &quot;div&quot;: &quot;&lt;div xmlns=\&quot;http://www.w3.org/1999/xhtml\&quot;&gt;&lt;div class=\&quot;hapiHeaderText\&quot;&gt;NuÃ±ez &lt;b&gt;KARLA &lt;/b&gt;&lt;/div&gt;&lt;table class=\&quot;hapiPropertyTable\&quot;&gt;&lt;tbody&gt;&lt;tr&gt;&lt;td&gt;Date of birth&lt;/td&gt;&lt;td&gt;&lt;span&gt;02 January 1980&lt;/span&gt;&lt;/td&gt;&lt;/tr&gt;&lt;/tbody&gt;&lt;/table&gt;&lt;/div&gt;&quot;<br>  },<br>  &quot;name&quot;: [ {<br>    &quot;family&quot;: &quot;Karla&quot;,<br>    &quot;given&quot;: [ &quot;NuÃ±ez&quot; ]<br>  } ],<br>  &quot;gender&quot;: &quot;female&quot;,<br>  &quot;birthDate&quot;: &quot;1980-01-02&quot;<br>}</pre><h3>DICOM: The language of images. And getting them places.</h3><p>DICOM is both a <strong>file format standard</strong> and a <strong>communication protocol</strong> for storing and transmitting medical images. DICOM ensures that all medical imaging equipment can speak the same language, much like what HL7 and FHIR does for text based data. A DICOM file might contain both pixel data and metadata regarding patient name, patient ID, age, etc..</p><p>DICOM files have the filetype .dcm</p><p>See? Just like a PDF. Simple. Well.. If PDF also had a communication protocol but you get the idea.</p><p>What about the DICOM protocol? DICOM uses a client-server model where systems communicate using the DICOM protocol over TCP/IP.</p><p>Key services include:</p><ul><li><strong>C-STORE</strong> — Upload (store) an image to another system (e.g., from scanner to PACS).</li><li><strong>C-FIND</strong> — Query for patient or study information.</li><li><strong>C-MOVE</strong> — Request image transfer to another device.</li><li><strong>C-GET</strong> — Retrieve images.</li><li><strong>C-ECHO</strong> — Test communication (similar to a “ping”).</li></ul><p>These sound like HTTP’s GET and POST but don&#39;t be fooled. Remember, HTTP works <strong>on top of TCP/IP</strong> and likewise, <strong>DICOM</strong> also works on top of TCP/IP. There&#39;s no HTTP in between DICOM and TCP/IP. DICOM procotol sends data in binary over TCP/IP.</p><p>Confused? Hold on a bit.</p><h4><strong>What are DICOM <em>service classes?</em></strong></h4><p>In DICOM, a <strong>Service Class</strong> defines a <strong>type of operation or workflow</strong> that one device (a <em>Service Class User</em>, SCU) can perform with another device (a <em>Service Class Provider</em>, SCP).</p><p>And each service class, has their own DICOM commands. For example,</p><p><strong>Service Class</strong> <strong>Purpose</strong> <strong>Typical Commands (Operations)</strong> <strong>Storage Service Class</strong> Store images or data objects on a remote system (e.g., scanner → PACS). C-STORE <strong>Query/Retrieve Service Class</strong> Search and retrieve patient/study/image data from PACS. C-FIND, C-GET, C-MOVE</p><p><strong>Okay those are the 4 elephants in the room. We’ll rapid-fire the rest since they’re not that hard to grasp</strong></p><h3>PACS: The image library</h3><p><strong>PACS</strong> is a <strong>software system</strong> (and storage server) used by hospitals to <strong>store, retrieve, distribute, and view medical images</strong>.</p><p>It uses the DICOM standard for all its image storage and communication. Its basically a glorified hard drive for DICOM images.</p><h3>EHR: The recordbook</h3><p>EHR stands for Electronic Health Record. It stores patient health information — diagnoses, medications, allergies, lab results, etc…</p><p>The EHR uses all the protocols listed above to achieve its task. It uses FHIR for communicating with servers to get the health data. It uses PACS to store the lab imaging. It uses DICOM to read, transmit and process the images.</p><h3>SNOMED CT: The dictionary</h3><p>This is just the global standard terminology for clinical terms (diseases, symptoms, findings)</p><p><strong>Go on now. Create a health-tech startup</strong></p><p>Just kidding. This is just the beginning of your journey. You’ll need to learn a lot more than this. But this’ll get you going. Run along now.</p><h3>Part 2: So You Want to Build a Health-Tech Startup (For Real This Time)</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5G7TxkHsr-5CHvcoPp2iBQ.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@nci?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">National Cancer Institute</a> on <a href="https://unsplash.com/photos/person-sitting-while-using-laptop-computer-and-green-stethoscope-near-NFvdKIhxYlU?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p>Alright, so now you know your FHIRs from your DICOMs and your PACS from your SNOMEDs. But theory only gets you so far. Let’s get our hands dirty. How do you <em>actually</em> build something in healthcare that works — and doesn’t accidentally summon the compliance gods?</p><h3>Step 1: Pick Your Battlefield — or in Plain English, Your Use Case</h3><p>The healthcare universe is <strong>massive</strong>. Don’t try to build an EHR from scratch on day one (that’s like deciding your first coding project will be “Windows 12”). Instead, find a focused niche:</p><ul><li>Appointment scheduling for clinics</li><li>Patient lab report viewers</li><li>Medical image sharing between doctors</li><li>AI-powered symptom checker</li></ul><p>Each of these uses a <em>subset</em> of the standards you just met. Knowing which ones matter saves you months of headache.</p><pre>| Use Case                    | Standards You’ll Likely Touch |<br>| --------------------------- | ----------------------------- |<br>| Patient data portal         | FHIR, OAuth2, HL7             |<br>| Image viewer                | DICOM, PACS                   |<br>| Lab result exchange         | LOINC, FHIR                   |<br>| Decision support            | SNOMED CT, FHIR               |<br>| Multi-hospital data sharing | IHE XDS.b, FHIR, ATNA         |</pre><h3>Step 2: Get a FHIR Server Up and Running</h3><p>FHIR is the heart of modern health data exchange. The easiest way to start is to use a ready-made open source FHIR server. The <strong>HAPI FHIR</strong> server (Java) is the go-to option.</p><p>You can:</p><ul><li>Spin it up locally with Docker:</li></ul><pre>docker run -p 8080:8080 hapiproject/hapi</pre><ul><li>Or use their public test instance at <a href="https://hapi.fhir.org/baseR4">https://hapi.fhir.org/baseR4</a></li></ul><p>Once it’s up, you can hit it with REST calls like:</p><pre>GET /Patient/123<br>POST /Observation</pre><p>Congratulations, you just spoke fluent healthcare.</p><p>If you’re a Python person, check out fhir.resources or fhirclient.They help you work with FHIR objects natively.</p><h3>Step 3: Handle Authentication Like an Adult</h3><p>If you’re building a health app, <strong>OAuth2 and OpenID Connect</strong> are your new best friends. Why? Because healthcare systems don’t hand over patient data to “random app from the internet dot com.”</p><p>Most modern FHIR servers integrate with <strong>SMART on FHIR</strong>, an authentication layer built on top of OAuth2. It standardizes how apps get access tokens and permissions (called <em>scopes</em>), like:</p><pre>launch/patient<br>patient/*.read<br>patient/Observation.write</pre><p>You’ll need to register your app, handle redirects, and exchange authorization codes — just like you would with Google or GitHub login. Except this time, you might be logging in through a hospital’s Keycloak instance.</p><h3>Step 4: Play Nice with Legacy Systems</h3><p>Let’s face it — most hospitals are still running ancient systems that look like they were built on Windows 95. They speak <strong>HL7 v2</strong>, not FHIR. You’ll need a <em>translator</em>.</p><p>Enter: <strong>Integration Engines.</strong></p><p>Tools like <strong>Mirth Connect</strong>, <strong>OpenHIM</strong>, or <strong>NextGen Connect</strong> can transform HL7 v2 messages into FHIR resources.</p><p>Think of them as the universal plug adapter between old and new healthcare tech:</p><pre>HL7 ADT^A01 (Patient Admitted)<br>   ↓<br>FHIR Patient + Encounter Resources</pre><p>With this bridge, your shiny modern web app can talk to hospital software from 2003 like they’re old buddies.</p><h3>Step 5: Respect the Rules — Privacy &amp; Compliance</h3><p>Healthcare isn’t like fintech or social media. Here, <strong>privacy is law</strong>, not preference.</p><p>Depending on where you operate:</p><ul><li>In the <strong>US</strong>, it’s <strong>HIPAA</strong></li><li>In the <strong>EU</strong>, it’s <strong>GDPR + MDR</strong></li><li>In <strong>Sri Lanka or India</strong>, it’s <strong>PDPA</strong>-style local laws Every country has its own flavor, but the gist is the same: <em>don’t leak patient data</em>.</li></ul><p>Use:</p><ul><li><strong>HTTPS everywhere</strong></li><li><strong>Encrypted storage</strong> (AES-256 or bust)</li><li><strong>Audit logs</strong> (FHIR even has an AuditEvent resource!)</li><li><strong>Access control</strong> (FHIR supports Consent resources for this)</li></ul><p>Trust is your most valuable currency in healthcare. Lose that, and you’re done.</p><h3>Step 6: Testing in the Wild</h3><p>Before you even think of touching real patient data, use <em>sandbox environments.</em></p><p>A few free ones:</p><ul><li><a href="https://hapi.fhir.org/">HAPI FHIR Public Test Server</a></li><li><a href="https://cloud.google.com/healthcare">Google Cloud Healthcare API Sandbox</a></li><li><a href="https://learn.microsoft.com/en-us/azure/healthcare-apis/">Microsoft Azure FHIR Server Sandbox</a></li><li><a href="https://inferno.healthit.gov/">Inferno Test Suite</a> — to check your FHIR compliance.</li></ul><p>These let you experiment safely and verify that your app’s FHIR implementation actually matches the spec (because trust me, hospitals will test that).</p><h3>Step 7: Ship, But Don’t Sink</h3><p>Once your prototype works, don’t just “move fast and deploy to production.” Remember: healthcare data is <em>regulated</em> data.</p><ul><li><strong>Pen test</strong> your system (and fix what breaks).</li><li><strong>Version your FHIR resources</strong> — healthcare data changes, and you’ll need audit trails.</li><li><strong>Monitor uptime</strong> — missing a few API calls is one thing, missing a patient’s allergy info is another.</li></ul><p>If your app survives this stage without catching fire, congrats. You’re no longer just “building in healthcare.” You’re <strong>building for human lives</strong>.</p><h3>Next Steps</h3><p>From here, you can explore:</p><ul><li><strong>LOINC</strong> for lab codes</li><li><strong>OMOP/CDISC</strong> for research data</li><li><strong>IHE ATNA</strong> for audit trails and security</li><li><strong>FHIR Subscriptions</strong> for event-driven updates</li><li><strong>Terminology Servers</strong> for SNOMED CT and friends</li></ul><p>But that’s for another day. For now, get your FHIR server running, hook up some dummy patients, and start building.</p><p>Healthcare is complicated — but with the right tools, you can <em>actually</em> make it better.</p><p>Till next time 👋</p><p>This article was also published on the <a href="https://blog.rotaractmora.org/modernizing-the-healthcare-enterprise-a-beginners-guide-for-startups">Rotaract Mora blog</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a9a4d223c648" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I only knew HTML and CSS. How I got my first website up and running in 3 weeks.]]></title>
            <link>https://medium.com/@limahaliyana/i-only-knew-html-and-css-how-i-got-my-first-website-up-and-running-in-3-weeks-baae72e070fb?source=rss-9020d25f0558------2</link>
            <guid isPermaLink="false">https://medium.com/p/baae72e070fb</guid>
            <category><![CDATA[django]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[backend]]></category>
            <category><![CDATA[full-stack]]></category>
            <category><![CDATA[programming]]></category>
            <dc:creator><![CDATA[Lasan Mahaliyana]]></dc:creator>
            <pubDate>Sat, 28 Aug 2021 01:30:40 GMT</pubDate>
            <atom:updated>2021-08-28T01:30:40.680Z</atom:updated>
            <content:encoded><![CDATA[<p>3 weeks + 15 hour days + enough caffeine to kill a water buffalo. That basically sums up my entire schedule.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*77SHHU8kQr7YoE3ZM7b0uQ.jpeg" /><figcaption>Credits : <a href="https://unsplash.com/@egle_sidaraviciute">Egle Sidaraviciute</a> on Unsplash</figcaption></figure><p>So basically… yeah I worked my ass off. But that’s to be expected from someone knowing only HTML and CSS right? There are two parts to web development. Front-end and back-end. Front end is where all the user interface lives.</p><blockquote>“Make the front-end pretty and go nuts on the back end.”</blockquote><p>That was my initial thought process. Well it did work for the most part. But… not so great in production and maintenance. We’ll get to that later.</p><h3>How the hell did this guy pull off a website with knowing only how to center a div?</h3><p>You could say I had an ace up my sleeve. No, I didn’t have an evil twin brother who happened to know the dark magic of server side programming. In fact, you could say that I’m pretty good at learning on the job. <a href="https://helpatyourself.herokuapp.com/">This website</a> was also my first project. How’s that for learning on the job? (I made that website as something to help my fellow students here in Sri Lanka at their studies. You can go check it out. Still has a lotta kinks to iron out though)</p><p><strong>In all fairness I did have a background in python, so that helped a lot too.</strong></p><p>So I was pretty confident I could dress my website in pretty clothes, but then I would have the clothes without the actual person. Which is stupid.</p><h3>Taking the leap and learning Django</h3><p>I first researched what first steps were necessary to get some traction. I found a decent guide that some Medium user had put together (Sorry, can’t remember the article itself) and used it as guidance. I found out that building everything from scratch would mean I would have to quit school (that ain’t happening) and the easiest way is to use a web framework. A web framework, as I soon found out, would have pre-built packages that would handle all of that. So I decided to throw in my lot with Django, a framework that was conveniently based on Python.</p><p>Django is a web framework designed for rapid development and deployment. It’s learning curve its basically not a curve at all. Its pretty linear all things considered. If you know you’re fair share of OOP, you’re pretty much set to go. It has a <strong><em>very </em></strong>extensive documentation <a href="https://docs.djangoproject.com/en/3.2/">over here</a>. I also followed this <a href="https://developer.mozilla.org/en-US/docs/Learn/Server-side/Django">helpful article from the good folks at mozilla</a>.</p><p>Turned out, programming the back end was not that hard at all. Ha! Gotcha there. Nope. It was grueling.</p><h3>What do you get when you mix a front-end dev with back-end frameworks and try to get a usable product?</h3><h4>Simple answer: Development Nightmare</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*29KzhfL8h85O-pllhs1zVA.jpeg" /><figcaption>(Think of suitable caption)</figcaption></figure><p>HTTP requests. Huh? User authentication. What now? POST and GET. You mean the post office?</p><p>That represents my 3 brain cells working together in harmony to understand what the heck I’d gotten myself into. Lots of protocols, lots of ways Django handles each one. Weird words getting thrown about like ORM and SQL queries.</p><p>If it was one massive problem, I’d have been fine. But this was death by papercuts. URL dispatcher kept routing people to wrong links. The half-baked django template language routinely kept driving me insane. And in the middle of testing, my database decided it had enough of my nonsense and went haywire. Which naturally forced me to rebuild the entire thing.</p><p>So, yeah. Not fun. I was working late into the night and waking up early the very next day, only to start up my laptop, pound my keyboard and bash my head into the monitor in frustration.</p><h4>Getting my shi*t together</h4><p>Within a two weeks or so of mish-mashing stackoverflow code, some other code that I mostly got off tutorials plus some that I actually created myself, I got myself the first django app. It was a far cry from what you see today, but progress was progress. I delved deep into the documentation (for once in my life) and even checked out some of the source code. This gave me a really good understanding of the fundamentals of the whole framework. It also made me appreciate the beauty and elegant design philosophy of the django project.</p><p>This gave me more confidence to leverage the immense capabilities of the framework and add more features and functionality.</p><h3>What do you mean “Deployment”?</h3><p>Okay, so I finished building the damn thing. Now, what?</p><p>When you develop a website, it is not accessible by anyone on the internet (Unless, you’re a network guru who actually knows how to host a website on a personal server. But that’s unsustainable and these days, stupid, unless you have very specific requirements). There are many many hosting services available on the internet. The problem is choosing the correct one.</p><p>I had just developed a python app. So I needed something that wasn’t too difficult to setup, while also supporting python out-of-the-box and quite possibly, free. I had two options.</p><ol><li><a href="http://www.pythonanywhere.com/">PythonAnywhere</a></li><li><a href="http://www.heroku.com/">Heroku</a></li></ol><p>PythonAnywhere had a really weird CPU time limit for free customers. You could only have 100s of CPU time. FOR A MONTH. I’d like to meet the genius who came up with that. Just setting up the environment took 10s of CPU time. Not to mention that for a total newbie like me, getting my files onto my server proved to be a real hassle itself.</p><p>In comparison, Heroku was much easier. Their servers can automatically deploy from git. Meaning that, you just push the files to the git repository, and heroku magically takes care of the rest. Their step by step guides on how to set them up is super useful. And their features for the free tier is really good. You get 650 dyno hours. (Which is heroku-speak for, “your server will be up for 650 hours”) plus and additional 350 hours if you add a credit card. <strong>Per month. </strong>That’s a bargain. Honestly.</p><h3>Since then…</h3><p>I have learnt a lot of stuff since then. I learn Javascript to spice things up in the UI. The homescreen right now, is as good as you can get. I totally revamped the entire design after realizing the old one was horse-crap. I’m also learning about websockets and the<a href="https://channels.readthedocs.io/en/stable/"> django channels project</a> to incorporate a chat application into my original. Fingers crossed.</p><p>And remember what I said about my approach “Make the front-end pretty and go nuts on the back-end” ? Yeah, it wasn’t really successful. It DID get the job done, but now, I set aside two entire days to properly refractor my codebase and to make it look less “frankenstein-ey”. Even so, I still see bits of redundant code from time to time.</p><p>Well… that just about sums up my journey into full-stack web dev. I never thought I’d be able to get a full on website up and running. (The left side of my brain is asking “What did it cost” from the right side. Right side: “Everything”). I guess that’s full stack web development for you guys. The thing with full-stack is that you have to be a one man army. Design, testing, implementation, deployment, maintenance, everything is managed by one behemoth of a man. On top of that add homework and classes, you have a nasty concoction (I’m a full-time student).</p><p><em>Whew, that was a nice little trip down memory lane. I hope this helps anyone who is trying to get into the web development sector. I have linked all the websites I used all across the article should you need them. Make sure to clap, subscribe, bunny-hop or all three if you want to see more like this.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=baae72e070fb" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Becoming a programmer while also being a full-time student. How?]]></title>
            <link>https://medium.com/@limahaliyana/becoming-a-programmer-while-also-being-a-full-time-student-how-6de123248cc4?source=rss-9020d25f0558------2</link>
            <guid isPermaLink="false">https://medium.com/p/6de123248cc4</guid>
            <category><![CDATA[student-life]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[students]]></category>
            <category><![CDATA[studying]]></category>
            <category><![CDATA[programming-tips]]></category>
            <dc:creator><![CDATA[Lasan Mahaliyana]]></dc:creator>
            <pubDate>Fri, 27 Aug 2021 01:15:20 GMT</pubDate>
            <atom:updated>2021-08-27T01:15:20.999Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*Hh0tZkacIlAObuNlz7eeAA.jpeg" /></figure><p>Becoming a programmer isn’t easy, especially with a full-time job or when you’re a student. But, with the right amount of work, it can be done.</p><p>So, how would you go about doing it? Well… the first step would be to,</p><h3><strong>1. Start off slow</strong></h3><p>It&#39;s easy to get into a very aggressive start and then burn out. You have other obligations too. Academic pressures such as exams, deadlines, and social obligations like parties and other general human stuff (Family meetings. ugh). If you’re like me, you love programming and get lost in a sea of code without noticing time fly by. So, it&#39;s important to set aside time for other activities.</p><p>What you want to do is <strong>space out your schedule. </strong>Don’t try to cram in 4-hour coding sessions in the middle of the night and re-enact the walking dead at school tomorrow.</p><h3>2. Be consistent</h3><p>Some days, you just don’t feel like doing it. Some days, you just don’t have the energy. But learning to program is like learning a language (probably because it IS a language). The only way you’re going to get better at it is by using it consistently. When you don’t use it that often, you soon forget about the little things like syntax, and proper coding practices.</p><p>Committing to a language is especially important. Don’t delude yourself by saying “I can learn both these languages” when you’re just starting out. Usually, when you’re a beginner you don’t have a very thorough knowledge of what goes where and how everything connects. Different programming languages use different methods (and vastly different syntax too, sometimes) to do the same thing, and getting the two mixed up can be very frustrating.</p><p>For example take this c++ code:</p><pre>#include &lt;iostream&gt;<br><br>using namespace std;<br><br>int main(int argc, const char * argv[])<br>{<br>    char yesno;<br><br>    cout &lt;&lt; &quot;Enter &#39;y&#39; for yes and &#39;n&#39; for no.&quot;;<br>    cin &gt;&gt; courseLevel;<br><br>    if (yesno == &quot;y&quot;)<br>    {<br>        cout &lt;&lt; &quot;You said something&quot;;<br>    }<br><br>    return 0;<br>}</pre><p>and this python code</p><pre>inp = input(“input ‘y’ for yes and ’n’ for no”)</pre><pre>if inp == “y”:</pre><pre>   print(“You said “ + “yes”)</pre><pre>if inp == “n”:</pre><pre>   print(“You said “ + “No”)</pre><p>The c++ code would give us an error as it&#39;s essentially comparing a character to a string. Python is smart enough to figure this out. But this error is just so bizarre for someone coming from python that it might take them <em>hours </em>to find the cause (sadly, speaking from experience).</p><h3><strong>3. Use the internet. Wisely</strong></h3><p>The internet is the biggest resource you can possibly have access to. There are virtually tens of thousands of tutorials on programming in various different languages. It is basically a well of knowledge that’ll never run dry. Use them properly to grow your knowledge and learn from the best. Let’s talk about some sites that are useful to new and budding programmers as well as veterans.</p><ol><li>Youtube (Duh)</li><li>Stackoverflow (For Copy-paste )</li><li>freeCodeCamp</li><li>W3Schools (Mostly web dev + python and c++)</li><li>Codecademy (For the fundamentals)</li></ol><p>Youtube, is hands down google’s gift to mankind (I know, I know, they didn’t invent it) and is filled top to bottom with great video tutorials on programming. And the best thing is, they’re all free. Although it might be free and congested with good programming content, I’ve found out that most videos don’t really get deep into the theoretical side of programming. For a beginner programmer, this should be more than enough though.</p><p>Stackoverflow is regarded as the holy grail of programmers. There are over <strong><em>50 million monthly active members </em></strong>contributing to the site, which is reason enough for its usefulness. People like ranting on about how toxic the community is but honestly, I’ve asked a few dumb questions there myself and the worst I’ve ever gotten is a link to the documentation.</p><p>freeCodeCamp is a coding bootcamp website that teaches you a wide variety of programming languages at no cost. You heard it right. Absolutely free. Although keep in mind that the entire course is grueling with more than 300 hours of programming involved. Not for the faint-hearted, but one of the most complete programming bootcamps on the internet.</p><p>If you’re into web dev and would love to learn the ins and outs of the internet, w3schools is worth checking out. It guides you through all the necessary things to build you’re own website including back-end and front-end programming.</p><p>Know nothing about programming? Codecademy is the place to go. Its not free, but they offer in depth lessons on the fundamentals making it one of the first places new programmers should check out. Its not free though.</p><h3><strong>4. Use multiple sources of knowledge</strong></h3><p>This one kind of blends in with my previous point. Don’t stick to just one source of knowledge. There might be some things that it does extremely well, and… not so well. If you’re a full time student leaning programming, chances are, you don’t even <strong><em>know</em></strong> what you have to learn. So you might miss out on some important concepts without even knowing it.</p><p>If you can, its always good to find a more experienced mentor who can guide you. It doesn’t have to be a professional programmer (though, more experienced the better), it could be a fellow student who’s gone down that same path and achieved some level of proficiency.</p><p>Don’t be afraid to spend some money. Courses on Udemy, Coursera can be pretty expensive, but dig around enough and you might find a good course on a discount. Programming is a vast field and paid courses guarantee you more material and higher quality content than free ones.</p><h3><strong>5. Do more projects.</strong></h3><p>This is one that’s especially useful. You need to get a feel for the development cycle of any app or website or whatever you’re building. You need real world experience of planning out a project, coding it, debugging it, testing and maintenance.</p><p>Sign up on a website like fiverr and charge people for your work. That is, after you’re confident that you can manage anything they throw at you. Start out on something slow and easy, like scripting or web scraping. This way, you get more experience working with clients and what the development process looks like, and gets paid for your efforts at the same time.</p><p>Learning a version control system like Git is also important as it familiarizes you with programming tools that professionals use.</p><h3><strong>6. Follow up with school</strong></h3><p>Most schools have cs curriculums that gets hands down with an easy programming language like python. Don’t skip those classes and try to integrate your own learning progress into your classwork. (Okay I admit, learning to open Excel isn’t helping anybody, but try to pay attention to programming).</p><p>Try to integrate school work into programming. Got an assignment to enter 100 excel rows? Make a script for it. Zoom chat disabled by teacher? Make a website with a chat room for your buddies (Speaking from experience kiddos. Don’t tell mom). Don’t view school work as a hindrance, but instead, try incorporate it into the learning process.</p><h3><strong>7. Practice. Practice. Practice</strong></h3><p>Do I even need to say this? Practice and consistency is the key to success, no matter which path you wander down. Practicing makes it easy to learn all the syntactical differences between languages and makes you code faster. Practice makes perfect (Well… in theory, yes, but even after 5 years of programming, I still have 200 browser tabs open. But it does help)</p><p>If you enjoyed this, make sure to… I dunno, do whatever you need do to see more content from me. Clap, subscribe, bunny-hop whatever, I dunno.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6de123248cc4" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>