Because the real meetup.com didn’t like my idea…

EventFetch is a demo, Android “no-root” application (jq script) for fetching structured data and displaying it to the user.

It is just an example file. The real meat is in the system of standards.

Structured data, to me, is the “Network’s HTML”. It is often human readable, we (programmers) send it back and forth over the network, whenever we need some event or something to be logged somewhere.

All it is, is a let of standards. Here is an abbreviated JSON blob for one of my websites:

<!-- Structured Data -->
<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "ProfessionalService",
  "@id": "https://meadowlarklowvoltage.com/#business",
  "name": "Meadowlark Low Voltage",
  "url": "https://meadowlarklowvoltage.com/",
  "telephone": "+1-610-707-9247",
}

Not much to it, is there…

Its literally just headings and fields sort of situation. Something =s something. Fantastic, Object-Oriented Programing I can understand at 2AM! (JSON is not OO LOL it was a joke).

EventFetch Structured Data Example

[ { "id": "it worked!", "url": "https://site.somwthing/demo-event", "author":
    "host@event.com", "published": "2026-09-14T12:00:00Z", "tags": [
      "event", "US", "PA", "12345", "20260920", "1900" ], "content": "EVENTFETCH demo\nevent in somewhere fun Planet 3"
  }
]

Lets “Pretty Print” it:

[
  {
    "id": "it worked!",
    "url": "https://site.somwthing/demo-event",
    "author": "host@event.com",
    "published": "2026-09-14T12:00:00Z",
    "tags": [
      "event",
      "US",
      "PA",
      "12345",
      "20260920",
      "1900"
    ],
    "content": "EVENTFETCH demo\nevent in somewhere fun Planet 3"
  }
]

I love structured data!

Really, I am a “web dev”. When a customer wants SEO, I make them structured data first, then we move into the fancy methods… Usually by the end of the “meta writing” as I call structured data creation. Usually, we have spent a lot of time and have positioned them to be vastly more competitive in the market than they were before. From an SEO perspective.

Simply, the world runs on structured data! We have all these applications all written by different people, often the only way they share anything is threw this sort of “third type of language”. Which is not a language at all, but AN AGREED-ON SET OF RULES AND EXPECATIONS. A protocol of sorts.

Federated Event – Structured Data Type

I herby propose, by the power vested in me by my laptop, that there be a standard “Event Structured Data Type”. So that across instances, everyone can pull events and compile lists.

I have created EventFetch as a demonstration of how small an application can be.

I have also specifically built in the door from plugins, because this is the real world, and it needs to be flexible. for “plugins”. So, if an instance or relay does not support the “baseline data structure that I cobbled together on my laptop”. You can simply redefine the fetcher with a plugin to properly phrase whatever source you would like. There is a Mastodon example.

THE CODE!!!

Was written by a robot. But is in shell! Read it as you would text!

It was written by a robot because that way a robot will be able to better fix it. This is just a demo; the expectation would be someone could use it to create their own. This one is very basic, all it does is run through the sources blobs and chunk them into files, all locally. Somone could write their own with pencil and paper or however you fancy. I used to try to write scripts before LLMs, it took HOURS!

First is the breakdown, then full thing.

Chunk by Chunk:

#!/data/data/com.termux/files/usr/bin/sh

SOURCE="$1"
PLUGIN="${2:-}"
COUNTRY="$3"
REGION="$4"
POSTAL="$5"
EVENT_DATE="$6"
EVENT_TIME="${7:-}"

OUT="${EVENTFETCH_OUT:-$HOME/eventfetch/events}"
MAX_BYTES="${EVENTFETCH_CHUNK_SIZE:-1048576}"

TMP="$HOME/.eventfetch"

NORMALIZED="$TMP/normalized.jsonl"
MATCHES="$TMP/matches.jsonl"
RECORD="$TMP/record.txt"
CHUNK="$TMP/chunk.log"

mkdir -p "$OUT"
mkdir -p "$TMP"

BASE="events-${COUNTRY}-${REGION}-${POSTAL}-${EVENT_DATE}"

if [ -n "$EVENT_TIME" ]; then
    BASE="${BASE}-${EVENT_TIME}"
fi

chunk_no=1
chunk_records=0
total_records=0

: > "$CHUNK"

if [ -n "$PLUGIN" ]; then
    "$HOME/eventfetch/plugins/$PLUGIN" "$SOURCE" > "$NORMALIZED"
else
    curl -s "$SOURCE" | jq -c '.[]' > "$NORMALIZED"
fi

jq -c \
    --arg country "$COUNTRY" \
    --arg region "$REGION" \
    --arg postal "$POSTAL" \
    --arg date "$EVENT_DATE" \
    --arg time "$EVENT_TIME" '

    ([.tags[]? | ascii_downcase]) as $t |

    select($t | index("event")) |
    select($t | index($country | ascii_downcase)) |
    select($t | index($region | ascii_downcase)) |
    select($t | index($postal | ascii_downcase)) |
    select($t | index($date | ascii_downcase)) |

    select(
        ($time == "")
        or
        ($t | index($time | ascii_downcase))
    )

' "$NORMALIZED" > "$MATCHES"

seal_chunk() {

    if [ "$chunk_records" -eq 0 ]; then
        return
    fi

    FILE=$(printf "%s/%s-%04d.log" "$OUT" "$BASE" "$chunk_no")

    mv "$CHUNK" "$FILE"

    BYTES=$(wc -c < "$FILE")

    echo "$(basename "$FILE")    ${BYTES} bytes    ${chunk_records} records"

    chunk_no=$((chunk_no + 1))
    chunk_records=0

    : > "$CHUNK"
}

stop_fetch() {
    echo
    echo "Stopping..."
    seal_chunk
    echo
    echo "$total_records records saved."
    exit
}

trap stop_fetch INT

while IFS= read -r POST
do

printf '%s\n' "$POST" |
jq -r '
    "========================================================================",
    "SOURCE: " + (.url // ""),
    "AUTHOR: " + (.author // ""),
    "PUBLISHED: " + (.published // ""),
    "TAGS: " + ([.tags[]?] | join(" ")),
    "",
    (.content // ""),
    "========================================================================",
    ""
' > "$RECORD"

RECORD_BYTES=$(wc -c < "$RECORD")
CHUNK_BYTES=$(wc -c < "$CHUNK")

if [ "$chunk_records" -gt 0 ] &&
   [ $((CHUNK_BYTES + RECORD_BYTES)) -gt "$MAX_BYTES" ]
then
    seal_chunk
fi

cat "$RECORD" >> "$CHUNK"

chunk_records=$((chunk_records + 1))
total_records=$((total_records + 1))

CHUNK_BYTES=$(wc -c < "$CHUNK")

if [ "$CHUNK_BYTES" -ge "$MAX_BYTES" ]; then
    seal_chunk
fi

done < "$MATCHES"

seal_chunk

echo
echo "Fetch complete."
echo "$total_records records saved."

All together:

#!/data/data/com.termux/files/usr/bin/sh

SOURCE="$1"
PLUGIN="${2:-}"
COUNTRY="$3"
REGION="$4"
POSTAL="$5"
EVENT_DATE="$6"
EVENT_TIME="${7:-}"

OUT="${EVENTFETCH_OUT:-$HOME/eventfetch/events}"
MAX_BYTES="${EVENTFETCH_CHUNK_SIZE:-1048576}"

TMP="$HOME/.eventfetch"

NORMALIZED="$TMP/normalized.jsonl"
MATCHES="$TMP/matches.jsonl"
RECORD="$TMP/record.txt"
CHUNK="$TMP/chunk.log"

mkdir -p "$OUT"
mkdir -p "$TMP"

BASE="events-${COUNTRY}-${REGION}-${POSTAL}-${EVENT_DATE}"

if [ -n "$EVENT_TIME" ]; then
    BASE="${BASE}-${EVENT_TIME}"
fi

chunk_no=1
chunk_records=0
total_records=0

: > "$CHUNK"

if [ -n "$PLUGIN" ]; then
    "$HOME/eventfetch/plugins/$PLUGIN" "$SOURCE" > "$NORMALIZED"
else
    curl -s "$SOURCE" | jq -c '.[]' > "$NORMALIZED"
fi

jq -c \
    --arg country "$COUNTRY" \
    --arg region "$REGION" \
    --arg postal "$POSTAL" \
    --arg date "$EVENT_DATE" \
    --arg time "$EVENT_TIME" '

    ([.tags[]? | ascii_downcase]) as $t |

    select($t | index("event")) |
    select($t | index($country | ascii_downcase)) |
    select($t | index($region | ascii_downcase)) |
    select($t | index($postal | ascii_downcase)) |
    select($t | index($date | ascii_downcase)) |

    select(
        ($time == "")
        or
        ($t | index($time | ascii_downcase))
    )

' "$NORMALIZED" > "$MATCHES"

seal_chunk() {

    if [ "$chunk_records" -eq 0 ]; then
        return
    fi

    FILE=$(printf "%s/%s-%04d.log" "$OUT" "$BASE" "$chunk_no")

    mv "$CHUNK" "$FILE"

    BYTES=$(wc -c < "$FILE")

    echo "$(basename "$FILE")    ${BYTES} bytes    ${chunk_records} records"

    chunk_no=$((chunk_no + 1))
    chunk_records=0

    : > "$CHUNK"
}

stop_fetch() {
    echo
    echo "Stopping..."
    seal_chunk
    echo
    echo "$total_records records saved."
    exit
}

trap stop_fetch INT

while IFS= read -r POST
do
    printf '%s\n' "$POST" |
    jq -r '
        "========================================================================",
        "SOURCE: " + (.url // ""),
        "AUTHOR: " + (.author // ""),
        "PUBLISHED: " + (.published // ""),
        "TAGS: " + ([.tags[]?] | join(" ")),
        "",
        (.content // ""),
        "========================================================================",
        ""
    ' > "$RECORD"

    RECORD_BYTES=$(wc -c < "$RECORD")
    CHUNK_BYTES=$(wc -c < "$CHUNK")

    if [ "$chunk_records" -gt 0 ] &&
       [ $((CHUNK_BYTES + RECORD_BYTES)) -gt "$MAX_BYTES" ]
    then
        seal_chunk
    fi

    cat "$RECORD" >> "$CHUNK"

    chunk_records=$((chunk_records + 1))
    total_records=$((total_records + 1))

    CHUNK_BYTES=$(wc -c < "$CHUNK")

    if [ "$CHUNK_BYTES" -ge "$MAX_BYTES" ]; then
        seal_chunk
    fi

done < "$MATCHES"

seal_chunk

echo
echo "Fetch complete."
echo "$total_records records saved."

An Archive to play with…

If you are running android, get Termux from the play store, you can copy that file to a tree you create (if you think my download has viruses haha). Or you can download it from the button.

You can change the top line and run it anywhere on Linux.

eventfetch/
├── fetch.sh
├── README.md
├── events/
└── plugins/
    └── mastodon

THERE IS A DEMO EVENT LIVE AT THIS DOMAIN at /eventfetch-demo.json (https://proe.whimm.ing/eventfetch-demo.json)

What to expect in Termux:

I don’t even want you to download it, or any of this to do anything…

I want someone else too. I want someone who is smarter and better and has more time than me. To spawn something like this. The efficiency we could gain as people if we learned tools like this.

To call that file an “application” is a stretch… But to me, it does an application like thing! It goes and gets data, brings it back to me. All from my phone. Without any bloat or ads or anything.

All it needs is other people to share the same freaking “UHHH OK SO ADDRESS IS SPELLED A-D-D-R-E-S-S AND YOU WRITE IT LIKE THIS”. That’s it, if we could just pull that off, every individual human would be able to use the internet to directly communicate with their peers. Without any sort of “platforms” or ads”

Again, screaming into a void, just an idea I had.

BRING ON THE AGE OF MICRO LINUX APPS FOR ALL!

Peter Roe