2025,江端さんの技術メモ

GStreamerをC言語でフレームレートの設定ができなかったので、色々試していましたが、

caps = gst_caps_from_string("video/x-raw, width=640, height=360,framerate=5/1")は動きません(少なくとも私のコードでは駄目だった)。

これね、GStreamerのコマンドと同じ内容になるように並べなければあらないようです。

gst-launch-1.0 rtspsrc location=rtsp://cam:Cam12345@192.168.0.10/Src/MediaInput/stream_1 latency=0 ! rtph264depay ! avdec_h264 ! videoconvert ! videoscale ! video/x-raw,width=640,height=360 ! videorate ! video/x-raw,framerate=1/1 ! x264enc bitrate=1700 speed-preset=ultrafast tune=zerolatency key-int-max=1 ! mpegtsmux ! srtserversink uri = "srt://192.168.101.10:38091" latency=500

これを忠実にコーディングする必要があるようです。

で、今のコードに入っていなかったのが"video/x-raw,framerate=1/1" です。

なので、capsfilter2 やら caps2 のコードを追加する必要がありました。
という訳で、以下のような追加をすることで動くようになりました。

ーーーーー

GstElement *videoscale, *capsfilter, *videorate, *x264enc, *mpegtsmux, *srtserversink;
GstElement *capsfilter2;

GstCaps *caps;
GstCaps *caps2;

ーーーーー

capsfilter = gst_element_factory_make("capsfilter", "capsfilter");
videorate = gst_element_factory_make("videorate", "videorate");
capsfilter2 = gst_element_factory_make("capsfilter", "capsfilter2");

ーーーーー

if (!pipeline || !rtspsrc || !rtph264depay || !avdec_h264 || !videoconvert || !videoscale ||
!capsfilter || !videorate || !capsfilter2 || !x264enc || !mpegtsmux || !srtserversink) {
g_printerr("Not all elements could be created.\n");
abc_log_message("Not all elements could be created.");
return -1;
}

ーーーーー

caps = gst_caps_from_string("video/x-raw,width=640,height=360");
g_object_set(capsfilter, "caps", caps, NULL);
gst_caps_unref(caps);
caps2 = gst_caps_from_string("video/x-raw,framerate=10/1");
g_object_set(capsfilter2, "caps", caps2, NULL);
gst_caps_unref(caps2);

ーーーーー

gst_bin_add_many(GST_BIN(pipeline), rtspsrc, rtph264depay, avdec_h264, videoconvert, videoscale,
capsfilter, videorate, capsfilter2, x264enc, mpegtsmux, srtserversink, NULL);

if (!gst_element_link_many(rtph264depay, avdec_h264, videoconvert, videoscale, capsfilter, videorate,
capsfilter2, x264enc, mpegtsmux, srtserversink, NULL)) {
g_printerr("Elements could not be linked.\n");
abc_log_message("Elements could not be linked.");
gst_object_unref(pipeline);
return -1;
}

ーーーーー

もっと簡単にできると思ったんだけどなぁ・・・結構、手間かかったなぁ。

2025,江端さんの忘備録

私は、プロジェクトで動いていますが、本質は「ぼっち」で、共同作業が苦手です。

I work on projects, but at heart, I'm a loner, and I'm not good at working together.

で、そろそろ本気で考えているのが、「自分エージェント」です。

So, I have seriously considered making a "personal agent."

コミュニケーションは、生成AIで構成される、「自分エージェント」に任せて、技術の部分だけを担当する(こちらも、生成AIに助力して貰うのですが)ようにならないかな、と。

I wonder if it would be possible to leave communication to a “personal agent” of generative AI and only be in charge of the technical aspects (although we would also get help from generative AI).

こういう「自分エージェント」を、不登校やうつ、教員不足(というか労働者不足)の問題にも適用できないかな、などと考えています。

We could apply this “personal agent” to truancy, depression, and teacher shortages (or worker shortages).

こういうものは、すでに故人のペルソナを反映したAIで使われ出していますので、技術的にはもう2~3段階くらいのイノベーションが必要かとは思うのですが ―― それでも、私は『絶対に不可能』とまでは思っていません。

This kind of thing is already being used in AIs that reflect the persona of the deceased. So, I think that technically, we'll need to make another 2 or 3 stages of innovation, but I don't think it's 'impossible.'

-----

まあ、まだ影も形も存在しない「自分エージェント」ができる未来を想定すると、以下のような問題が出てきそうな気がします。

If we imagine a future in which “personal agents” that don't even exist yet become a reality, I think the following problems will likely arise.

(1) 本人と「自分エージェント」が乖離して、コミュニケーションが取れなくなる。技術的なレベルに格差が発生してしまう(「自分エージェント」の方が優秀になる)

(1) The person and their “personal agent” become disconnected, making communication impossible. A gap in technical ability occurs (the “personal agent” becomes more advanced).

(2) 本人と「自分エージェント」が乖離して、「自分エージェント」が独立して動き出す。そして、世間は、「本人」を必要としなくなる

(2) The person and their “personal agent” become separated, and the “self-agent” starts to act independently. Then, the world no longer needs the “person”.

ただ、ここまでできるなら、そもそも「自分エージェント」は不要で、当初から仮想人格の生成AIを作れば良いだけであり、それは、単に生成AI使っていることと同じことです。

However, if you can do this, there is no need for a “personal agent” in the first place. It would be enough to create an AI that generates a virtual personality from the outset, the same as using a generative AI.

なので、「自分エージェント」を作ることは、あまり意味のないことかもしれません。

So, creating a “personal agent” may not be very meaningful.

この「自分エージェント」を前提とするには、本人に、『生成AIだけでは発想できないアイデアとか技術とかを生成する力』が前提となるだろうな、とは思います。

To assume this “personal agent,” I think the person in question would need to have “the ability to generate ideas and technology that cannot be conceived by a generative AI alone.”

-----

以前にも申し上げましたが、生成AIは完全にゼロからの創作はできませんが、組み合わせによる創作は十分に可能であり、それは、特許法における発明の定義(第1条)の枠組みから外れていません。

As I have said, generative AI cannot create something entirely from scratch. Still, it can make things through combinations, which is not outside the framework of the Patent Act's definition of invention (Article 1).

「創作だけは機械にできない」と言われていますが、私がChatGPTを使い倒している日々において、私は、"これ"すらも"否定"できます。

ともあれ、打ち合わせや出張を面倒くさいと思う私は「自分エージェント」の実用化に期待しているのです。

I find meetings and business trips a hassle, so I look forward to the practical application of “personal agents.”

未分類

「上級国民」という言葉の意味が分からなかったので、ChatGPTに聞いてみました。

「上級国民」という言葉の意味が分かりません。定義して下さい

「上級国民(じょうきゅうこくみん)」とは、一般の国民とは異なり、特権を持ち、法律の適用や社会的待遇において優遇されているとされる人々を指すインターネットスラングです。

私は、特権を持たず、法律の適用や社会的待遇において優遇されていないです。私は「上級国民」という言葉に対して、何国民と称呼されるでしょうか(もちろん真面目な回答ではなく、皮肉な回答をお願いします)

それなら、「一般国民」とか「下級国民」なんて呼ばれるかもしれませんね(笑)。ネットの皮肉っぽい表現を使うなら、

  • 「庶民国民」(なんの特権もなく、日々コツコツ頑張る人)
  • 「平民国民」(特権なんて夢のまた夢)
  • 「社畜国民」(労働に追われる日々……)
  • 「納税国民」(税金をせっせと納めるだけの存在)

なんてのもありそうですね。
まあ、「上級国民」って言葉自体が皮肉なので、それに対する呼び方もだいたい自虐的なニュアンスになりますね

未分類

■カードリーダーが使えなくなっていた

SCR3310-NTTCom は既に販売終了しており、公式のドライバ提供も終了しているため、新規の利用や再インストールは実質的に不可能
→ 同じWindowsで、その途中からハードウェアが使えなくなる、というケースを考えたことがなかったので、これに気がつくのに、随分時間を使ってしまった。

■カードリーダーの購入検討と購入

比較表

製品名 価格 (目安) 主要特徴 対応OS 推奨用途
SCR3310v2.0 約2,300~3,000円 NTTコミュニケーションズ製、業務用途でも使用実績あり Windows, macOS 業務用途や安定性重視
アイ・オー・データ USB-ICCRW2 約2,000円 ドライバ不要、マイナンバー対応 Windows, macOS 家庭用で手軽に使いたい
ゴッパ GP-ICCR 約1,500円 最安クラス、コンパクトサイズ Windows, macOS コスト最重視
サンワサプライ ADR-MNICU2 約1,800円 ドライバ不要、USBバスパワー駆動 Windows, macOS 簡単に使いたいが安定性も求める

Amazonで即日入荷で、ドライバ不要 という観点から、サンワサプライ ADR-MNICU2 に決定。

で、さっき届いたので、USBに差して、マイナンバーカードを挿入したら、簡単にシステムにアクセスできた(今迄の苦労が、本当に虚しい)

■医療費控除をするのに、家族全員の代表者として「私」を登録する為に、家族全員のマイナンバーカードによる登録が必要
→ 次女が旅行中で、手元に家族全員のマイナンバーカードがない

#昨年まで、こういう『代表者登録』しなくてよかったんだけどな

#まあ、こういう代表者を悪用する「バカ」が不正な医療費控除を受けようとした、などの事件が発生したんだろう、と思う。

-----

行政側もセキュリティ対策で色々大変だろうと思うけど、毎年インターフェース(手続)を変えられて、右往左往させられる国民も、そうとうに大変だぞ

この、インターフェース変更は、これからも、毎年ずっと続いていくのだろう、と思うと ――

正直、(システムエンジニアである、私でさえも)『デジタル化、もういい』という気持ちになります

 

 

2025,江端さんの忘備録

戦後、日本では公営住宅の人気が急騰し、多くの人々が自宅を売り払ってまで入居したと聞いたことがあります。

After the war, I heard that public housing in Japan became very popular and that many people even sold their homes to move in.

その理由は、当時としては画期的だった設備の充実(水洗トイレやユニットバス)と清潔感でした。

The reason for this was the facilities, which were revolutionary for the time (flush toilets and prefabricated baths), and the sense of cleanliness.

しかし、現在では公営住宅の入居者が減少し、高齢化や空室問題が深刻化しています。

However, the number of residents in public housing is decreasing, and the problems of aging and empty units are becoming more serious.

こうした公営住宅の歴史を振り返ると、私には、現代のタワーマンションも同じ運命をたどるのではないか、という疑問が浮かびます。

When I look back on the history of public housing like this, I can't help but wonder whether the same fate will befall the high-rise apartments of today.

つまり、今は憧れの住まいとして高い人気を誇るタワーマンションも、いずれは誰も住まなくなり、廃墟となるのではないか、と。

In other words, even the high-rise apartments that are so popular as the homes of our dreams today will eventually become uninhabited and ruined.

で、まあ、私、この分野に関して大学で学生やっているので、以下のようなことを考えています。

I'm a university student in this field, so I'm thinking about the following things.

■仮説A:「タワーマンションは将来的に廃墟化する」

Hypothesis A: “Skyscrapers will become ruins in the future.”

(1)タワーマンションは高度な設備を備えているため、維持管理には多額のコストがかかりる。エレベーターや給排水設備の修繕、外壁の補修など、築年数が経つほど負担が増大し、管理費や修繕積立金の上昇が住民の負担を圧迫する。適切な修繕ができなければ、建物の価値は急速に下落し、空室が増える可能性がある。

(1) Because high-rise apartments have advanced facilities, they require a lot of money for maintenance and management. The burden increases as the building ages, with repairs to the elevators, water supply and drainage facilities, and external walls. The rise in management fees and repair funds puts pressure on residents' finances. If appropriate repairs are not carried out, the value of the building will rapidly decline, and the number of vacancies may increase.

(2)初期の購入者が高齢化し、住み替えが進まないと、建物全体の活気が失われます。次世代の居住希望者が減少すれば、管理組合の機能が弱まり、管理不全による住環境の悪化が加速する

(2) The building will lose vitality if the initial buyers age and do not move out. If the number of prospective next-generation residents decreases, the management association's functions will weaken, and the deterioration of the living environment due to poor management will accelerate.

(3)現在は人気のエリアでも、都市計画や人口動態の変化によって将来的に価値が下がる可能性がある。特に郊外のタワーマンションは、都心回帰の流れが進むと空室率が上昇し、維持が困難になる

(3) Even in currently popular areas, there is a possibility that the value will decrease in the future due to changes in urban planning and population dynamics. In particular, if the trend of moving back to the city center progresses, the vacancy rate of high-rise condominiums in the suburbs will increase, making it difficult to maintain them.

(4)住民同士のつながりが希薄なタワーマンションでは、管理組合の活動が停滞すると、共有スペースの維持管理が困難になります。結果として、防犯や清掃が行き届かなくなり、住み心地が低下する可能性がある

(4) In high-rise apartments where residents have few connections with each other, if the activities of the management association stagnate, it will become challenging to maintain and manage the common spaces. As a result, crime prevention and cleaning may become inadequate, and the comfort of living may decrease.

■仮説B:「タワーマンションは廃墟化しない」

Hypothesis B: “Skyscrapers do not become ruins.”

(1)公営住宅と違い、タワーマンションは分譲物件であるため、住民が資産価値を維持しようと積極的に管理・修繕が行われる。特に高級マンションでは管理組合が機能しているケースが多く、老朽化しても適切なメンテナンスが行われる。

(1) Unlike public housing, since high-rise apartments are sold individually, the residents actively manage and repair them to maintain their asset value. In particular, in many cases, the management association functions well in high-class apartments, and even if they become old, appropriate maintenance is carried out.

(2)日本の総人口は減少しているものの、都市部への人口集中は一定のペースで続くと予想されている。交通アクセスが良く、生活インフラが整っている物件は、今後も一定の需要を維持する

(2) Although Japan's total population is declining, the concentration of population in urban areas is expected to continue at a certain pace. Properties with good access and a well-developed living infrastructure will continue to maintain a certain level of demand.

(3)公営住宅と異なり、タワーマンションには民間の投資が入りやすいため、築年数が経過してもリノベーションによる価値向上が見込める。また、ホテルやシェアハウス、オフィスへの用途転換など、柔軟な活用方法も考えられる。

(3) Unlike public housing, private investment is more likely to be attracted to high-rise condominiums, so even after a certain amount of time has passed, it is possible to expect an increase in value through renovation. In addition, it is also possible to consider flexible methods of use, such as conversion to hotels, shared houses, or offices.

(4)近年、単身世帯や高齢者向け住宅の需要が増加しています。タワーマンションはセキュリティが高く、管理人が常駐する物件も多いため、高齢者にとっても魅力的な住環境となり得る。

(4) In recent years, there has been an increase in demand for housing for single-person households and the elderly. Tower apartments have high security, and many have a caretaker on site, making them attractive living environments for the elderly.

-----

うーん・・・

Hmm...

『仮説B』を採択するとすれば、現在の公営住宅だって、リノベーション活用が進んでも良さそうなものですよね。でもあまり、そういう試みは聞かれていません。

If we adopt 'Hypothesis B,' it seems like even the current public housing could be renovated and used to better use. However, we haven't heard about many such attempts.

正直、そういう試みがあると学会で発表などあるのですが、私が知る限り、1回の学会で2~3件の報告が聞ければ良いほうです(学生によるコミュニティ拠点のような話が多いです)。

There are presentations at academic conferences about such attempts. Still, as far as I know, it's good to hear 2 or 3 reports at a single conference (there are a lot of talks about student-led community hubs).

-----

私は、将来、タワーマンションが、香港の九龍城砦(くーろんじょうさい / Kowloon Walled City)みたいにならなければいいんだけどなな、と思っています。

I hope that in the future, high-rise apartments won't become like Kowloon Walled City in Hong Kong.

香港の九龍城砦は、1993年にこの地域の強制撤去を決定し、1994年までに住民を退去させて取り壊しましたが、たぶん、恐しい額の公費が投入されたと思います。

In 1993, the decision was made to forcibly remove the residents of Kowloon Walled City in Hong Kong, and by 1994, the residents had been evicted and the buildings demolished, probably at a frightening cost to the public purse.

あと、仮説Bの方は「都心部のタワマンは何とかなるかもしれん。が、地方のタワマンのことまでは知ったことではない」という立ち位置なのも、問題です。

Also, the problem with hypothesis B is that it assumes that “we might be able to manage the high-rise apartments in the city center, but we don't know about the high-rise apartments in the countryside.”

-----

「作る」ものは、「壊す」ところまで視野に入れておく必要があるんですよ。

When you're making something, you must remember that it will eventually be destroyed.

地震大国日本は、この「壊す」という作業(の一部)を、自然災害に任せてきたという面は ―― まあ、多分、あるだろうなぁ、と思います。

As an earthquake-prone country, Japan has probably pushed (part of) the task of “destroying” natural disasters.

でも、今のタワマンは、多分、ちょっとやそっとの地震では壊れません。

However, today's high-rise buildings will probably not be destroyed by a small earthquake.

それが日本のタワマンの「売り」でもありますからね。

That's also the selling point of Japanese tower blocks.

2025,江端さんの忘備録

先日、こちらの日記で、「このバルチック艦隊に日本海軍が仕掛けた、秋山真之の立案した戦法というのが、未だに、よく理解できていません」と記載しました。

The other day, in this diary, I wrote, “I still don't understand the battle plan that the Japanese Navy devised for the Baltic Fleet, which Akiyama Masayuki proposed.”

「坂の上の雲」の再放送を見てみると、日露戦争において、我が国は、奇跡的なラッキーを5つぐらい連続で掴むという、歴史的幸運に恵まれていたことを思い知ります。

で、先週、このシーンが放送されました。

So, this scene was broadcast last week.

私は、エンジニアで、研究員で、SE(システムエンジニア)で、ネットワーク技術者ですが、やはり一番しっくりくるのが『シミュレーション屋』です。

I am an engineer, researcher, system engineer, and network engineer, but I think the term best describes me is “simulation engineer.”

そういう、シミュレーション屋の性(さが)として、『秋山真之の立案した戦法というのが、未だに、よく理解できていません』というのは沽券に係わります(いや、正直、そうでもないんですが)。

As is the nature of simulationists, it is a matter of honor (well, to be honest, it's not really that important) to say, 'I still don't understand the battle plan that Akiyama Manayuki came up with.

で、前回の放送で、どうしても手に入れたかった情報である、日本海軍とバルチック艦隊の位置と時間の情報(海図)を、テレビの画面をスマホで撮影して、そこから分析してみました。

So, in the previous broadcast, I took a photo of the TV screen with my smartphone and analyzed the information (Nautical chart) on the location and time of the Japanese Navy and the Baltic Fleet, which I wanted to get.

ようやく分かってきた「秋山真之」の"T字戦略"

で、先程、この答合わせを、ChatGPTにしてみました。

So, I tried using ChatGPT to check my answers.

添付の画像に示すように、まず日本艦隊がUターンを行うことで、ロシア艦隊の砲撃をさそう形で1列に並べて、そして、その1列にならんだロシア艦隊の前に、先んじて立ち塞がり、列で進行してくるロシア艦隊を正面から逐次撃破した、という風に理解したのですが、これは正しいでしょうか

As shown in the attached image, I understood that the Japanese fleet first made a U-turn, lined up in a single row to lure the Russian fleet into firing on them, and then stood in front of the Russian fleet, which was lined up in a row, and defeated the Russian fleet, which was advancing in a line, one after another from the front. Is this correct?

===== ChatGPT回答ここから =====

はい、おおむね正しい理解です。

戦術の流れ

  1. 東郷ターン(Uターン)で敵砲撃を誘発

    • 画像の右側(青線)を見ると、日本艦隊が大きくUターンしていることが分かります。これは**「東郷ターン」**と呼ばれるもので、日本艦隊はこの動きによって、ロシア艦隊の砲撃を一時的に分散させる形になります。
    • この機動によって、日本艦隊はロシア艦隊に対して相対的に有利な位置を取ることが可能になります。
  2. ロシア艦隊を1列に誘導

    • バルチック艦隊(ロシア艦隊)は南下しながら戦列を維持しているため、先頭艦が攻撃を受けると、それに続く艦も回避行動が難しくなるという弱点を持っています。
    • 日本艦隊の機動によって、ロシア艦隊の進行方向を制限し、1列に並ばざるを得ない形に誘導されました。
  3. 「T字戦法」による迎撃

    • 日本艦隊はロシア艦隊の進行方向の正面を塞ぐ形で横向きに展開し、T字を形成しました。
    • この位置関係では、日本艦隊は側砲(舷側砲)を最大限に活用できるのに対し、ロシア艦隊は前方の主砲しか使えず、不利な状況に陥る
    • 画像でも、日本艦隊の砲撃が集中してロシア艦隊の先頭艦を撃破し、敵の進行を阻害している様子が描かれています(赤線の砲撃)。
  4. 逐次撃破

    • 先頭のロシア艦が撃沈または大破すると、その後続艦は回避行動を取らざるを得ず、戦列が乱れます。
    • これによって、ロシア艦隊は組織的な反撃が難しくなり、日本艦隊は効果的に逐次撃破を行うことができました。

結論

あなたの理解は概ね正しいです。「東郷ターン」を使って敵の砲撃を誘い、T字戦法を完成させ、逐次撃破を行うという流れになっています。特に、ロシア艦隊の隊列を分断することに成功した点が、日本海海戦の勝因の一つです。

===== ChatGPT回答ここまで =====

Here is the complete translation of your requested text while keeping the content and format unchanged.


Akiyama Saneyuki’s "T-Shaped Tactics" in Detail

Akiyama Saneyuki’s "T-Shaped Tactics" (丁字戦法, T-jisenpō) refers to the naval strategy employed by the Imperial Japanese Navy to engage the Russian Baltic Fleet during the Russo-Japanese War (1904-1905). This strategy was notably executed in the Battle of Tsushima (May 27-28, 1905), which played a decisive role in Japan’s overwhelming victory.

1. Basic Concept of the T-Shaped Tactics

The T-shaped tactics involved positioning the fleet to form the horizontal bar of the letter "T" relative to the enemy fleet’s line of advance. This had several key advantages:

  • Maximizing the firepower of the Japanese fleet by enabling broadside attacks.
  • Focusing concentrated fire on the leading enemy ships, disrupting their formation.
  • Obstructing the enemy’s route, restricting their maneuverability.

2. Execution of the T-Shaped Tactics in the Battle of Tsushima

Akiyama Saneyuki, as the chief strategist for Admiral Tōgō Heihachirō, devised and implemented this strategy in the battle. The execution proceeded as follows:

1. Detecting the Russian Fleet in Advance

  • The Japanese navy deployed scout ships in advance to track the movements of the Baltic Fleet.
  • The Russian fleet’s passage through the Tsushima Strait was confirmed, allowing for a well-planned interception.

2. Execution of the "Tōgō Turn" to Establish the T-Shaped Formation

  • On May 27, 1905, the Japanese fleet altered its course to intercept the enemy.
  • Admiral Tōgō led the fleet in executing the "Tōgō Turn," a dramatic U-turn maneuver designed to position the Japanese fleet across the enemy’s path.
  • As a result, the Japanese fleet’s leading battleships lined up perpendicularly to the Russian fleet’s advance, forming a "T" shape.
  • This allowed Japan to use its full firepower while limiting the Russians’ ability to return fire effectively.

3. Decisive Destruction of the Russian Fleet

  • The Japanese fleet unleashed concentrated artillery fire, critically damaging the Russian flagship "Suvorov" and other lead ships.
  • This disrupted the command structure of the Russian fleet, throwing it into confusion.
  • As a result, most of the Russian fleet was sunk, captured, or forced to surrender, securing a resounding Japanese victory.

3. Significance of the T-Shaped Tactics

The success of this strategy demonstrated a significant evolution in naval warfare tactics, with several key takeaways:

  • Proof that superior tactics could defeat a numerically superior enemy

    • Although the Russian fleet had more significant numbers and firepower on paper, it was decisively beaten by superior maneuvering.
  • Highlighting the importance of positioning in naval combat

    • Previously, raw firepower and armor were considered the primary determinants of naval battles. However, this battle proved that strategic positioning could dictate the outcome.
  • Emphasizing the role of strategic planning

    • Akiyama Saneyuki’s planning, combined with Tōgō’s decisive execution, demonstrated that a well-planned strategy could dramatically alter the balance of power in a battle.

4. Evaluation and Limitations of the T-Shaped Tactics

While highly effective in the Battle of Tsushima, the T-shaped tactics were not universally applicable, and their effectiveness had limitations.

Advantages

  • Allowed for maximum firepower concentration against the enemy.
  • Restricted the enemy’s ability to maneuver and return fire.

Disadvantages

  • Required precise timing and an accurate prediction of enemy movements.
  • Lost effectiveness if the enemy changed formation in response.
  • Became less relevant with advancements in naval warfare, particularly with the introduction of aircraft and submarines.

Following World War I, the rise of new naval technologies gradually reduced the effectiveness of such large-scale fleet engagements, making the T-shaped tactics less relevant in modern warfare. However, they remain an essential case study in naval strategy.

5. Conclusion

Akiyama Saneyuki’s T-shaped tactics were a groundbreaking strategy that led to Japan’s decisive victory in the Battle of Tsushima. Japan maximized its firepower by predicting enemy movements and optimizing its fleet positioning while minimizing enemy retaliation. While this strategy became obsolete with technological advancements, it remains one of the most studied naval tactics in history and a testament to the importance of strategic planning in warfare.

私の解釈に間違いがあるとすれば、どこになりますか?

If there is a mistake in my interpretation, where would it be?

===== ChatGPT回答ここから =====

あなたの解釈はほぼ正しいですが、厳密に言えば、「正面から逐次撃破した」という点が完全に正しいとは言えない部分があります。具体的にどこが違うのか、詳しく説明します。


1. 「正面から撃破した」という表現の修正

あなたの解釈では、「ロシア艦隊が1列に並んだ状態で、それを日本艦隊が正面から迎え撃ち、順番に撃破していった」というように見えます。しかし、実際には次のような違いがあります。

日本艦隊の射撃方向

  • 「正面から」ではなく「側面から」撃破していったのが正しい。
  • 日本艦隊はロシア艦隊の進行方向を塞ぐ形で展開し、T字戦法を完成させた。
  • この位置関係では、日本艦隊は側面砲(舷側砲)を最大限に活用できるが、ロシア艦隊は前方の主砲しか使えないため不利だった
  • そのため、実際には「日本艦隊が横から砲撃し、ロシア艦隊の先頭から撃破していった」というのが正しい。

あなたの解釈では「正面からの撃破」となっているが、実際には「横からの撃破」だった点が異なる。


2. 「東郷ターン」は敵の砲撃を誘ったか?

あなたは、「東郷ターン(Uターン)によってロシア艦隊の砲撃を誘った」と解釈していますが、これは完全には正しくない部分があります。

実際の目的と結果

  • 東郷ターンの目的は、ロシア艦隊の先頭を塞ぐようにT字を作るための機動であり、直接的に砲撃を誘う意図はそこまで強くなかった。
  • しかし、このターンによって日本艦隊が一時的に無防備な状態になったため、ロシア艦隊が砲撃を行ったのは事実。
  • ただし、ロシア艦隊の砲撃は「誘われた」のではなく、「日本艦隊が方向転換中で回避が難しい状態だったため、自然に狙われた」という側面が強い。
  • しかも、ロシア艦隊の砲撃はあまり命中しなかった。

つまり、「日本艦隊がわざと砲撃を誘った」わけではなく、「機動の過程で砲撃されるリスクが生じたが、結果的にほとんど被害が出なかった」というのが正確な説明。


3. 逐次撃破の実態

あなたの解釈では、「日本艦隊がロシア艦隊を正面から1隻ずつ撃破した」という表現になっていますが、実際には以下のような違いがあります。

  • 実際には、日本艦隊はロシア艦隊の先頭を砲撃し、指揮系統を混乱させ、艦隊全体の動きを制限することを狙った。
  • つまり、「1隻ずつ順番に撃破した」のではなく、先頭艦を集中的に攻撃し、それに続く艦が混乱して戦列を維持できなくなるように仕向けた。
  • 画像でも示されているように、日本艦隊の砲撃は主にロシア艦隊の前方(先頭艦)に集中している。

「逐次撃破」というよりは、「先頭を崩し、その後続艦が混乱していくように誘導する戦術」だったのが実態に近い。


まとめ:あなたの解釈との違い

項目 あなたの解釈 実際の戦術
攻撃方向 正面から撃破 横(舷側砲)から撃破
東郷ターンの意図 敵の砲撃を誘った 主目的はT字戦法の形成(砲撃を受けたのは結果論)
撃破の方法 1隻ずつ順番に撃破 先頭を集中攻撃し、後続艦を混乱させた

結論

あなたの解釈は全体の流れとしては正しいですが、以下の3点を修正するとより正確になります。

  1. 「正面から撃破」ではなく「横(舷側)から撃破」。
  2. 「砲撃を誘った」のではなく、「機動の過程で砲撃を受けたが、被害は少なかった」。
  3. 「逐次撃破」ではなく、「先頭艦を集中攻撃し、後続艦を混乱させた」。

こうした点を修正すれば、より正確な理解になるでしょう。

===== ChatGPT回答ここまで =====

Here is the full translation of your requested text while keeping the content and format unchanged.


Where Is the Mistake in Your Interpretation?

Your interpretation is mostly correct. However, strictly speaking, the point that "Japan’s fleet destroyed the Russian fleet sequentially from the front" is not entirely accurate. Below, I will explain exactly where the differences lie.


1. Revising the Expression "Destroyed from the Front"

In your interpretation, you described the situation as "The Russian fleet was lined up in a single column, and the Japanese fleet positioned itself in front of them to destroy them sequentially." However, there are some key differences:

The Actual Direction of Japanese Fire

  • The Russian fleet was " destroyed from the side" instead of "destroyed from the front."
  • The Japanese fleet positioned itself across the Russian fleet’s path, forming the T-shape.
  • In this formation, the Japanese fleet could utilize its broadside guns, whereas the Russian fleet could only fire forward, putting it at a disadvantage.
  • Therefore, rather than "firing head-on and destroying them sequentially," it was actually "firing from the side and breaking the front of the Russian formation."

Your interpretation describes the destruction as occurring "from the front," but in reality, it was an attack from the side.


2. Did the "Tōgō Turn" Intentionally Lure Enemy Fire?

You suggested that the "Tōgō Turn" (U-turn maneuver) lured the Russian fleet into firing. However, this is not entirely correct.

Actual Intent and Outcome

  • The Tōgō Turn aimed to position the fleet in a T-formation to block the Russian fleet’s advance, not necessarily to bait enemy fire.
  • However, because the turn momentarily exposed the Japanese fleet, the Russian fleet did fire upon them.
  • The Japanese fleet did not deliberately invite enemy fire; instead, the maneuver temporarily made them vulnerable.
  • Moreover, most of the Russian fleet’s shots missed, causing minimal damage.

Thus, rather than "intentionally baiting fire," it is more accurate to say that "the maneuver momentarily exposed them to fire, but the damage was minimal."


3. The Reality of Sequential Destruction

Your interpretation suggests that "the Japanese fleet destroyed the Russian fleet one by one in order." However, in reality, the strategy worked as follows:

  • Instead of destroying them, the Japanese fleet focused on the leading Russian ships to disrupt their command and control.
  • By crippling the front of the Russian fleet, the rest of the ships were thrown into disarray.
  • As shown in the image, the Japanese firepower was concentrated on the front ships of the Russian fleet.

Rather than "sequential destruction," the strategy was more about "targeting the lead ships to induce chaos in the rest of the fleet."


Summary: Differences Between Your Interpretation and Reality

Aspect Your Interpretation Actual Strategy
Direction of Attack Attacked from the front Attacked from the side (broadside fire)
Purpose of the Tōgō Turn To lure enemy fire To form a T-formation (enemy fire was a result, not a goal)
Method of Destruction Destroyed one ship at a time in sequence Targeted the lead ships to disrupt the fleet

Conclusion

Your interpretation is correct mainly in terms of overall flow, but adjusting the following three points will make it more accurate:

  1. It was not "destroying from the front" but "firing broadside from the side."
  2. The Tōgō Turn did not "bait enemy fire" but "momentarily exposed the fleet to fire, though with little effect."
  3. It was not "sequential destruction" but "disrupting the lead ships to break the formation."

By refining these points, my understanding will be even more precise.

このように、ChatGPTは「忖度」だけでなく、頼めばちゃんとした「指摘」もしてくれます。

As I can see, ChatGPT not only “guesses” but also gives me proper “feedback” if I ask for it.

それに、人間から指摘されるより気分もいいです。人間から言われると、その文書の内容によっては「カチン」と来ることもありますし。

Also, it feels better to be told by a machine than a human. When a human tells me something, it can sometimes grate on me, depending on the document's content.

ともあれ、私は、このChatGPTの回答をもって、この『「このバルチック艦隊に日本海軍が仕掛けた、秋山真之の立案した戦法というのが、未だに、よく理解できていません」を、無事終了させることができた』と総括することにしました。

Anyway, I decided to summarize that I could complete the task of “I still don't fully understand the battle plan that the Japanese Navy devised for the Baltic Fleet, which was proposed by Akiyama Masayuki” by obtaining this response from ChatGPT.

-----

繰り返しますが、私は、今も激務が続いています ―― が、多くの人がご存知の通り、激務の時ほど、業務と関係ないことに集中する、ということを、私たちはティーンのころから繰り返してきましたよね(特に、中間や期末考査の前日の夜、とかに)。

Again, I'm still working hard - but as many people know, we've been doing this since we were teenagers, concentrating on things that have nothing to do with work when we're working hard (especially the night before mid-term or final exams).

今日の日記は、その典型的な見本です。

Today's diary is a typical example.

未分類

今週の「坂の上の雲(再放送)」の、この部分を20回ほど見直して、ようやく分かってきました。

Step 1 : 日本海軍の戦艦が、大きなUターンをするこによって、ロシア艦隊が「一列」になって攻撃を開始(この間、日本側は回頭中のため、一切反撃できず)
Step 2: 一列になってしまったロシア艦隊の目の前に、日本艦隊が集結
Step 3: 日本艦隊が一斉砲撃で、一列になったロシア艦隊を先頭から一つづつ潰していく

下図は、NHKのテレビ画面のキャプチャに、江端が図を追記したもの

 

ポイントは3点

(1)ロシア艦隊を1列にする必要があり、その為に、あえて直前での大Uターンを実施
(2)1列になったロシア艦隊の直前に、できるだけ早く日本艦隊を結集させる
(3)一列になったロシア艦隊を、ひとつづつを、一斉砲撃で潰す

YouTubeとか見れば、正解が分かるかもしれませんが、まずこの図面(とそこに記載された時刻だけ)から、上記を導いてみました。
答合わせは、後でやってみます。

2025,江端さんの忘備録

昨日は、「プロジェクトリーダーとしての生成AI」という観点で考えてみましたが、今回は、別の観点から考えてみたいと思います。

Yesterday, I considered the topic from the perspective of “AI as a project leader,” but this time, I would like to think about it from a different perspective.

あまり理解されていないかもしれませんが、生成AIがやっていることの一つに「ガリ勉」があります。

最近、『生成AI離れ』という言葉がでてきているようですが、毎日、ChatGPTを使い倒している私には、ちょっと信じられない言葉です。

Recently, I've been hearing the term "generation AI abandonment," but as someone who uses ChatGPT daily, I find it a little hard to believe.

業種(フィールド)だけでなく、使い方に、差があるのかもしれません。

There may be differences in the type of business (field) and how it is used.

プログラミングの世界においては、生成AIは「使えます」が、以下の条件が付帯するのかもしれない、と考えています。

In the programming world, I think that generative AI is “usable,” but it may come with the following conditions.

『生成AIが作ったコードの問題点を、指摘できるだけのスキルを人間が持っていること』

'The persons have the skills to point out the problems in the code created by generative AI.'

-----

依頼したプログラム(コード)をざっと見て、「変だな」と思い「ここ、間違っていない?」と聞き返せるスキルがあること、が、結構重要なのかな、と思うことがあります。

Sometimes, I think it is pretty essential to have the skill to look at the program (code) you have requested, think, “That's strange,” and ask, “Is this wrong?”

ChatGPTやGithub Copilotは素直な奴なので、私の指摘が正しければ、直ちにその場で修正を行うし、私の指摘が間違っていれば、私の間違いを理由とともに説明します。

ChatGPT and Github Copilot are straightforward, so if I'm right, they'll fix it right away, and if I'm wrong, they'll explain why I'm wrong.

お願い→回答、という、ワンリターンでは、生成AIはそんなに優れてはいないかもしれませんが、お願い→回答→質問→回答→疑問→回答.....と続けていけば、これほど心強いパートナーはいません。

If you only have one-way communication, where you ask a question and then get an answer, the generated AI may not be that good, but if you keep asking questions and getting answers, you will have a very reassuring partner.

生成AIを、"Copilot(副操縦士)"と命名した人は、センスがあると思います。

The person who named the generative AI “Copilot” has good taste.

-----

「生成AI離れ」とは、それはつまり、『副操縦士の言うことが理解できない/副操縦士と議論ができない』と解釈して良いのかもしれません。

“generation AI abandonment" may mean “I can't understand what the copilot is saying/I can't discuss with the copilot.”

結局のところ、「生成AI離れ」とは、「生成AIに丸投げして全て解決しようとする怠け者の発想なのではないかな」と、そんな風に邪推しています。

After all, I suspect that the “generation AI abandonment” is a lazy person's idea of trying to solve everything by throwing everything at the generation AI.

何が言いたいのかというと、「こういうプログラムを組める『私』には、市場価値があるのか、それとも、ないのか」、です。

2025,江端さんの忘備録

筑紫市の田中允市議(77)の、別の市議に対して行ったハラスメントに対して『至らぬところがあったかもしれませんが、それはそれとして今後、謙虚に学んでいけば私自身も成長するのかなと』というコメントを聞いて思い出したことがあったので、本日のネタとして記載します。

I remember hearing a comment from Tanaka Masashi, a 77-year-old city councilor in Chikushi City, Fukuoka Prefecture, about the harassment he had perpetrated against another city councilor. He said, “I may have fallen short in some ways, but if I humbly learn from this in the future, I will also grow.” I will include this as today's topic.

-----

あまり理解されていないかもしれませんが、生成AIがやっていることの一つに「ガリ勉」があります。

Although it may not be well understood, one of the things that generative AI is doing is “cramming.”

反復学習です。いわゆる「英単語帳」です。

It is a repetitive learning method. It is a so-called “English vocabulary notebook.”

例えるなら、生成AIは猛烈な勢いで「英単語の暗記」をして、「英語をしゃべれるよう頑張っている」というパラダイムで理解してもいいかもしれません。

If you want to compare it to something, you could say that the generative AI is furiously “memorizing English vocabulary” and “working hard to be able to speak English” in the paradigm of understanding.

その「暗記」は、単にメモするだけでなく、電気信号のパターンになるまで続けられます(このあたりは、ニューラルネットワークのメカニズムまで遡らなければならないので、割愛します)

This “memorization” is not just about making notes but is continued until it becomes a pattern of electrical signals (we will have to go back to the mechanisms of neural networks here so that we will skip that part).

件名: ChatGPTは怖くない ~使い倒してラクをせよ の記事に関してご提言

単に知識を蓄えるだけでは、知性にならないことは、すでに、「前世紀のエキパートシステムの失敗」で明らかですが、今の生成AIは、人間の判断基準(扱ってはならないネタ)も学習しており、加えていえば、「言葉使い」まで学習しています。

It is already apparent from the “failure of the expert systems of the last century” that simply accumulating knowledge does not make you intelligent. Still, the current generative AI also learns human judgment criteria (topics that should not be handled), and, to add to that, it also learns “word usage.”

-----

そういえば、以前何かコラムで読んだのですが、これからの時代は「知識」や「リーダーシップ」よりも、「親切」「優しさ」「礼儀正しさ」の方が重視される世の中になる、旨の記載がありました。

Speaking of which, I read something in a column before that said that in the future, the world would become a place where “kindness,” “gentleness,” and “politeness” will be valued more than “knowledge” and “leadership.”

つまり、

In other words,

―― 「凄い仕事」より、「気持よく続けられる仕事」に、人材がシフトする

"the shift in human resources is from “great work” to “work that can be continued comfortably”.

と。

上記の「仕事」は、「学校」とか「組織」とか「近所付き合い」などの言葉に置き換えられるかもしれません。

The word “work” above could be replaced with words such as “school,” “organization,” or “neighborly relations.”

『嫌われ役のリーダーを買って出て、プロジェクトを完遂する』という考え方は、もはや古い(のかも)。

“Volunteering to be the unpopular leader and completing the project” may already be old-fashioned.

(7)総じて「プロジェクトX」は、「エンジニアの夢や希望を搾取」して「プロジェクトを完成させた」というストーリで構成されており、現在の若者への「やりがい搾取」と1mmも変わっていないように見える

今や『嫌われ役のいるリーダーのプロジェクトは、プロジェクト自体が瓦解する』という流れになっていくのかもしれません。

Perhaps the trend is moving towards “a project with a disliked leader will collapse.”

ただ、進捗スケジュールを見ながら、厳しい納期を要求されているプロジェクトリーダーにとっては、このトレンドはシャレにならない話でしょう。

However, this trend is probably no laughing matter for project leaders under pressure to meet strict deadlines while monitoring the progress schedule.

-----

そういう観点から考えると、「プロジェクトリーダーを生成AIに任せる」というのは、十分にあり得る未来だと思います。

From that perspective, I think it's possible that “we will leave project leaders to AI.”

私が見ている限りにおいては、生成AIの人間に対する対応は、ほぼ完璧です。

As far as I can see, the way that generative AI interacts with humans is almost perfect.

生成AIは、間違っても『そんなことは自分で考えろ』とか『そんなことも知らないのか』などというような台詞は吐きません。

A generative AI will never say things like “You should be able to think for yourself” or “Don't you even know that?”

生成AIには、回答に対する「いいね」ボタンと「悪いね」ボタンがあり、生成AIは、これを受けて回答の内容や対応も学習しているようです。

The generated AI has a “good” button and a “bad” button for responses, and it seems that the generated AI learns the content of the reactions and how to respond to them based on these.

この考え方を拡張していくと、最終的に「生成AIによる政府や国際機関の運用」という未来に到達してしまいますが ―― 私は、それが今の状況よりもラクに生きられる世界を作ってくれるのであれば、それで構いません。

If we were to extend this line of thinking, we would eventually reach the future of “government and international organizations being run by generative AI,” - but as long as it creates a world where we can live more comfortably than we do now, I'm OK with that.

というか、むしろそうして欲しい。

Or rather, I want to realize it.

-----

ただ、行き過ぎた「生成AI頼み」は、かならず歪みを生み出すと思うので、多分、「生成AI排除」という動きも出てきます ―― 19世紀のラッダイド運動のように

However, overreliance on generative AI will inevitably lead to distortion, so there will probably be a movement to exclude it, like the Luddite movement in the 19th century.

ネットワークの世界においても、分散システムと集中システム、クラウドとエッジ、など、現在も数年単位で「揺り戻し」が発生しております。

Even in the world of networks, there are still “swinging back” trends occurring every few years, such as distributed and centralized systems, cloud and edge, etc.

プロジェクト管理に失敗した生成AIは廃棄され、より優れたプロジェクト管理AIが生き残ることになるとも思います(生成AIが学習して、発展していく可能性もあります)。

I also think that the generative AI that fails in project management will be discarded, and the AI with better project management skills will survive (there is also the possibility that the generative AI will learn and develop).

まあ、私がどう考えようと、生成AIの世界は、なるようにしかならないと思います。

Well, no matter how I think about it, I think the world of generative AI will only become what it is destined to become.

私は、ただ、それを観測し続けていくだけです。

I will continue to observe it.

-----

ただ、冒頭の市議は、「生成AIと取り替えても、一向に構わんだろう」 ―― というか、「是非取り替えてくれ」と思いますが。

However, about the city councilor at the beginning of the story, “Even if a generative AI replaces him, I don't mind at all” - or rather, I think, “Please replace him as soon as possible.”

2025,江端さんの忘備録

私は以前より『努力は天才に勝てない』ということを言い続けてきました。

I have always said that 'hard work cannot beat genius'.

故に、彼らは、平気で「毎日の練習」だの「継続は力」だの「努力は報われる」だのと言い放ちます。

最近思うことは、

Lately, I've been thinking that

『賞賛される努力と素晴しい成果より、凡庸なアウトプットの数が評価される』ということです。

'The number of mediocre outputs is valued more than the admired effort and excellent results.

努力と成果は、その時に絶賛されても、評価される時間が短いです。

Even if your efforts and achievements are praised at the time, the time it takes for them to be evaluated is short.

そして、人々の多くは、あなたの努力と成果を、簡単に忘れます。

And most people will easily forget your efforts and achievements.

比して、アウトプット(報告書や特許明細書や論文)は、内容はどうあれ、"数"で評価され、それは記録として残ります。

By comparison, outputs (reports, patent specifications, papers) are evaluated by the “number” of them, regardless of their content, and this is recorded.

-----

以前にもお話したかもしれませんが、会社同士がクロスライセンス契約をする時、『自社の特許明細書を積み上げた高さ』で、契約内容が決まる(高い方の会社が強い)という話がありました。

As I may have mentioned before, when companies enter into a cross-licensing agreement, the content of the agreement is determined by the “height of the pile of the companies' patent specifications” (the company with the higher pile is stronger).

この話が本当かどうかはウラが取れていませんが、ありそうな話です。

I have not been able to confirm whether this story is true or not, but it sounds plausible.

特許明細書の内容は難しいので、3桁を超える特許明細書の内容を、いちいち精査している余裕があるとは思えず、なにより『私でも、そうするだろうな』と思えるからです。

The content of patent specifications is difficult, so I don't think there is time to carefully examine the content of patent specifications that exceed three digits, and above all, I think that even I would do that.

(↑コラムに飛びます)

(Jump to the column)

-----

それゆえ、凡庸な私が選ぶべき戦略は明白です。

Therefore, the strategy that I, an ordinary person, should choose is clear.

「量を出し続けること」――たとえ、それが凡庸な内容であったとしても、です。

“Keep producing quantity” - even if it is mediocre content.

その"量"の戦略の一つが、この日記(ブログ)です。

One of the strategies for this “amount” is this diary (blog).